Accurate and dynamic monitoring of wheat phenotypes is essential for breeding decision-making and crop management. However, RGB image-based phenotyping still suffers from expensive pixel-level annotation, unstable organ-level segmentation across growth stages, and limited multi-trait extraction under complex field conditions. To address these issues, a high-throughput phenotyping framework (WheatScoper) was proposed, enabling organ-level segmentation and plot-level multi-trait extraction. To reduce annotation cost, a structure-aware geometry-assisted annotation (SAGA) algorithm was developed, yielding an approximately 7.6-fold improvement in annotation efficiency over fully manual annotation. To enable efficient organ-level segmentation, a lightweight semantic segmentation network (WheatScopeNet) was developed by integrating parallel hybrid spatial modeling with cross-scale feature fusion. On the held-out test set from the same site and growing season, WheatScopeNet achieved an mIoU of 0.869 and an mDice of 0.930. Leveraging the segmentation results, an automated system was established to extract 41 multi-view image-derived traits (I-traits) across four core phenotypic dimensions. The extracted I-traits supported the estimation of eight manually measured agronomic traits, with R 2 values ranging from 0.477 to 0.697. Notably, the correlations between stay-green-related-traits and yield varied distinctly with viewing-position. Only upper side-view indicators remained significantly correlated with yield, with Side-up final GPAR showing the strongest association, whereas top-view GPAR-derived indicators showed weak associations. Finally, a web-based platform integrating cascaded inference, segmentation visualization, and automatic I-trait extraction was developed. The platform provides an end-to-end solution for field wheat phenotyping and supports breeding decision-making and crop management.
Why it matches plant phenotyping methodsRGB画像から小器官を分割し、多数の植物形質を自動抽出・推定する手法とプラットフォームが研究の中心であるため、植物フェノタイピング手法として明確に採用。
abstracta high-throughput phenotyping framework (WheatScoper) was proposed, enabling organ-level segmentation and plot-level multi-trait extraction.
Plant phenotyping relevance match · UnverifiedbioRxiv · OpenAlex · Europe PMC · Crossref · checked 15 Sept 2026
Reliable plant segmentation in high-throughput phenotyping must transfer across species and imaging conditions without repeated model tuning or extensive reannotation. We compare three segmentation strategies using images from Oak Ridge National Laboratory's Advanced Plant Phenotyping Laboratory: (i) fixed color-based thresholding, (ii) supervised U-Nets trained from scratch, and (iii) pretrained vision transformers fine-tuned for binary segmentation. Models were evaluated on a held-out test set and a generalization set that comprised unseen species. On the held-out test set, thresholding, the best U-Net, and the best vision transformer achieved mean Dice scores of 58.3, 96.6, and 97.3, respectively. On the generalization set, the corresponding Dice scores were 56.5, 86.2, and 95.7. Thresholding remained effective on some datasets but failed when plant appearance changed. Supervised U-Net training resolved within-distribution errors but failed to generalize to novel species and backgrounds. Pretrained vision transformers consistently produced high-accuracy segmentations across the evaluated species, views, soil backgrounds, and tray types. These results benchmark the practical progression from fixed rules to task-specific supervision and pretrained visual representations for controlled-environment plant phenotyping.
Why it matches plant phenotyping methods植物フェノタイピングにおける画像セグメンテーション手法を比較・検証し、異なる種や撮像条件への汎化性能をベンチマークしているため、方法が研究の中心である。
abstractReliable plant segmentation in high-throughput phenotyping must transfer across species and imaging conditions without repeated model tuning or extensive reannotation.
High-throughput acquisition of crop phenotypic information is one of the key technologies for achieving intelligent facility agriculture and precision breeding. Traditional phenotypic data collection methods suffer from low efficiency and strong subjectivity, making it difficult to achieve multi-scale continuous monitoring and meet the demands of modern research and production. This paper systematically reviews the technological framework and development trajectory of optical sensing technology-driven phenotypic platforms for facility crops. First, starting from optical sensing technologies, a comparative analysis highlights the advantages and limitations of RGB, multi-/hyperspectral, thermal infrared, and LiDAR sensors in phenotypic perception. Second, the characteristics and applicable scenarios of stationary, rail-mounted, mobile robot, and unmanned aerial vehicle (UAV) platform architectures are summarized. Furthermore, the evolution of phenotypic data processing methods is examined, focusing on the shift from traditional feature engineering to deep learning-driven approaches. Finally, key challenges such as multimodal data fusion, system cost, and real-time performance are discussed, along with the future direction of phenotypic platforms toward intelligent closed-loop decision-making systems. This article systematically reviews the facility agriculture phenotyping platforms driven by optical sensing technology, and also incorporates representative research progress in field phenotyping studies. These advances provide transferable sensing technologies, methodological frameworks, and platform design concepts that can facilitate the development of phenotyping platforms for controlled-environment agriculture.
Why it matches plant phenotyping methods施設作物の光学センシング型ハイスループット表現型解析プラットフォームを体系的にレビューしており、センサー、プラットフォーム構成、データ処理を中心に扱うため、方法論レビューとして明確に適格です。
abstractThis paper systematically reviews the technological framework and development trajectory of optical sensing technology-driven phenotypic platforms for facility crops.
Abstract Purpose Agrivoltaic vineyards show strong spatio-temporal variability in canopy shading, but field methods to quantify panel-induced shading at canopy scale remain limited. Shading is a key factor because it affects plant physiological and morphological traits, with potential consequences for yield and production quality. This study developed a near-surface time-lapse RGB imaging approach to derive temporally explicit shading metrics in an agrivoltaic vineyard of Vitis vinifera cv. Falanghina in Southern Italy. Methods Two representative vine positions beneath the photovoltaic structure were monitored: Agrivoltaic Shade (AVS), with greater exposure to panel-induced shading, and Agrivoltaic Light (AVL), with lower exposure. Image-based canopy shading percentage was calculated through a dedicated processing workflow and integrated with radiometric and physiological measurements, including continuous photosynthetically active radiation (PAR), canopy-level spectral photon flux measurements, photosynthetic photon flux density (PPFD), band-specific photon flux densities, red:far-red ratio (R:FR), stomatal conductance (gₛ), and leaf temperature. PAR measurements beneath the panels were compared with a full-sun control area. Results AVS showed significantly higher shading than AVL (76.14% vs 39.45%, p Conclusion The proposed workflow offers a low-cost, non-destructive tool to quantify shading dynamics and support site-specific assessment of crop microenvironments in agrivoltaic systems. The approach provides crop-relevant information for precision monitoring and management of spatially heterogeneous light conditions across different crop species. Impact The data provided in this manuscript enable the quantification of in-season photovoltaic-induced canopy shading dynamics in an agrivoltaic vineyard using proximal RGB time-lapse imaging and crop-level radiometric measurements. These metrics reflect the spatial and temporal variability of light availability within the vineyard and support site-specific assessment of crop microenvironments and precision management of agrivoltaic systems.
Why it matches plant phenotyping methodsブドウ樹冠の遮光状態をRGBタイムラプス画像から定量化する手法を開発し、専用処理ワークフローと実測値で評価しており、植物フェノタイピング手法が中心である。
abstractThis study developed a near-surface time-lapse RGB imaging approach to derive temporally explicit shading metrics in an agrivoltaic vineyard of Vitis vinifera cv. Falanghina in Southern Italy.
Plant diseases, being a subject of interdisciplinary research, significantly reduce crop yield, quality, and economic returns, while the misidentification of pathogens often leads to ineffective treatments and may harm beneficial organisms and ecosystems. This work develops an approach for robust visual classification of plant diseases under limited and heterogeneous data based on multi-scale fractal texture descriptors integrated into a convolutional neural network. The proposed method employs wavelet transform modulus maxima to extract two complementary fractal characteristics, local fractal dimension and singularity spectrum width, from leaf images at several spatial scales. These descriptors form multi-channel fractal maps fed into a fractal attention module (FAM) inserted after the third stage of a ResNet-50 architecture. The FAM learns to emphasize spatial regions where fractal properties are most discriminative, while a parallel branch encodes global fractal statistics into an auxiliary vector combined with backbone features at the final classification layer. Experiments are conducted on a large heterogeneous collection of 11 public plant disease datasets under 5-shot, 50-shot, and full-scale training regimes. The fractal-augmented model raises classification accuracy from 57.06% to 67.73% on 5 shots and from 80.81% to 86.11% on 50 shots, red outperforming the plain ResNet-50 in these settings, converges within 1–2 epochs versus 25–40, and shows markedly better resilience to color distortions, random occlusions, and grayscale conversion in most cases. The generated attention maps provide spatially explicit explanations of the model’s decisions, increasing transparency for practical use. The proposed approach demonstrates that fractal analysis, embedded as a modulating signal inside a deep network, can serve as an efficient and interpretable inductive bias, which is particularly valuable under data scarcity and noisy agricultural imagery.
Why it matches plant phenotyping methods植物葉画像から病害状態を推定するフラクタル特徴抽出とCNN手法の開発・評価が中心であり、植物病害フェノタイピング手法に該当する。
abstractThis work develops an approach for robust visual classification of plant diseases under limited and heterogeneous data based on multi-scale fractal texture descriptors integrated into a convolutional neural network.
This study assessed the integration of Unmanned Aerial Vehicles (UAVs) and satellite images (Sentinel-1 and Sentinel-2) in advanced machine-learning techniques to monitor the ABV of quinoa crops (Jacha Grano variety) across the Bolivian Altiplano. The proposed method follows a two-step procedure. First, UAV RGB images were used in photogrammetric and deep-learning (Convolutional Neural Networks-CNN) models to estimate reference quinoa ABVs at a 10 m spatial resolution from the crop canopy 3D model and classification, respectively. Secondly, several spectral and polarization/texture indices derived from Sentinel-2 and -1 images were integrated into three decision-tree-based machine-learning models (Random Forest-RF, Gradient Boosting-GB, eXtreme Gradient Boosting-XGB), and one CNN-based machine-learning model to estimate ABV. Additionally, a Stacking Model (STM) build on top of the three decision-tree-based models was considered for comparison. Model evaluation was also performed in a two-step approach. First, a 10-fold cross-validation strategy was used to highlight ABV sensitivity to Sentinel-2 and Sentinel-1 alone and in combination. Secondly, a Leave-One-Plot-Out Cross-Validation (LOPOCV) strategy was used to avoid autocorrelation between the training and evaluation dataset and therefore provided more insight into ABV mapping potential. The results showed that the combination of Sentinel-1 and Sentinel-2 features in the CNN model achieved the best predictive performance with R2 and RMSE values of 0.64 and 0.39 m3 ∙ 100 m−2, respectively. These findings highlight the potential of integrating multi-source information in advanced artificial intelligence algorithms for quinoa ABV monitoring, offering new insights toward the identification of sustainable practices across remote regions with complex socio-economic contexts.
Why it matches plant phenotyping methodsUAV画像・衛星センサー・機械学習を統合し、キノアの地上部バイオマス体積という植物形質を推定する手法を開発・比較検証しており、フェノタイピング手法が中心である。
abstractUAV RGB images were used in photogrammetric and deep-learning (Convolutional Neural Networks-CNN) models to estimate reference quinoa ABVs
Abstract Plant diseases substantially reduce global crop yields, and cotton production is particularly vulnerable to field-acquired variability in symptom appearance, background clutter, and illumination changes that limit the reliability and scalability of expert visual inspection. This study aimed to develop an accurate, computationally efficient, and explainable framework for real-time cotton leaf disease recognition that is suitable for deployment on resource-constrained edge devices. Using the SAR-CLD-2024 dataset (322 RGB images captured under natural agricultural conditions across seven categories, including healthy and diseased leaves), images were preprocessed via resizing and normalization and augmented online in the training set (random rotations, flips, brightness/contrast adjustments, and random cropping). An EfficientNet-B3 backbone initialized with ImageNet-pretrained weights was fine-tuned using categorical cross-entropy loss and Adam optimization, with early stopping, checkpointing, regularization, and a fixed-seed 70/15/15 train–validation–test partition to enhance reproducibility and reduce leakage. Performance was evaluated on an independent test set using accuracy, precision, recall, F1-score, MCC, balanced accuracy, Cohen’s kappa, confusion matrix, multi-class ROC/AUC, and precision–recall analysis, alongside computational benchmarking (parameters, FLOPs, memory, and inference latency) and comparative experiments against contemporary CNN, lightweight, and transformer-based models. The model showed stable convergence over 30 epochs with a small training–validation gap, predominantly correct predictions with limited confusion among visually similar classes, consistently high precision–recall behavior under moderate class imbalance, and stable performance across repeated runs with low variability and a tight confidence interval. Grad-CAM heatmaps localized necrotic lesions, discoloration, and infected tissues while largely ignoring background, and failure cases were associated with early-stage symptoms, occlusion, shadows, and inter-class similarity. Overall, the framework provides a reproducible, interpretable, and efficient solution for cotton leaf disease classification with practical implications for trustworthy, low-latency, on-device decision support in precision agriculture.
Why it matches plant phenotyping methods綿葉の病徴を画像から分類する深層学習手法の開発が中心で、独立テスト、比較評価、計算性能評価、Grad-CAMによる病徴局在化を実施しているため、植物病害フェノタイピング手法に該当する。
abstractThis study aimed to develop an accurate, computationally efficient, and explainable framework for real-time cotton leaf disease recognition
Reproduction assets foundThe paper's Data Availability statement explicitly names the SAR-CLD-2024 cotton leaf dataset used for all experiments as publicly available on Kaggle with a direct URL. No author analysis code or trained model checkpoint is deposited.Dataset · publicntribute to the development of fully automated, scalable, and real-time smart agriculture
systems.
Declaration
Funding
Datta Meghe Institute of Higher Education and Research Wardha, Maharashtra, India
Data Availability: The SAR-CLD-2024 cotton leaf dataset used in this study is publicly
available through the Kaggle platform at: https://www.kaggle.com/datasets/pantho12/sar-cld-2024-dataset-for-cotton
This dataset includes annotated images of various cotton leaf diseases collected under diverse
environmental conditions. All data utilized in this work are freely accessible, and the data
processing methodology has been described in detail to facilitate reproducibility.
Conflict of interest The aOpen asset ↗Kaggle · SAR-CLD-2024pdf-raw-page:32 lines:1-38Plant phenotyping relevance match · UnverifiedCrossref · checked 14 Sept 2026
Published1 Sept 2026International Journal of Applied Earth Observation and Geoinformation
Accurate plant disease detection remains challenging when using single-modality data, which fails to capture comprehensive disease-related features. However, many existing studies rely on pixel-level classification or prior plant segmentation and lack explicit modeling of cross-modal interactions, limiting their ability to distinguish between healthy and diseased plants. This study presents a spatial–spectral fusion deep learning framework (S2-PDD) that fuses uncrewed aerial vehicles (UAV)-based RGB imagery and hyperspectral-derived feature representations for plant-level detection of potato diseases (blackleg and Potato virus Y (PVY)), employing early fusion (modalities combined at the input stage) and middle fusion (features integrated at intermediate stages within the model backbone) strategies. The multimodal fusion models were compared against single-modal models and an existing S2ADet model. Model performance, assessed through five-fold cross-validation, demonstrated that multimodal models integrating RGB and vegetation index features achieved the highest mAPs of 86.65 ± 1.71 (%; E-RV model) and 85.74 ± 1.96 (M-RV), respectively. These mAPs were higher than those of all single-modal models, including the RGB-only (83.21 ± 1.46) and hyperspectral-only models (PCA features: 79.71 ± 1.45; vegetation index features: 85.31 ± 2.36). They also exceeded mAPs of multimodal models combining RGB with PCA features (early fusion: 83.00 ± 2.81; middle fusion: 83.11 ± 2.46; S2ADet: 84.04 ± 2.73), regardless of the fusion strategy. The superior performance highlights that vegetation index features provide strong class separability compared to other hyperspectral representations. The proposed models achieved strong plant-level detection performance, with AP of 78.65 ± 4.14 (E-RV) and 77.40 ± 3.46 (M-RV) for blackleg disease, as well as 84.82 ± 3.59 (E-RV) and 83.28 ± 3.86 (M-RV) for PVY. These results demonstrate the potential of UAV-based multimodal sensing for disease monitoring in cropping systems. A potato plant disease detection dataset was constructed and made publicly available, containing paired RGB and hyperspectral image tiles with bounding box annotations. The code is available at https://github.com/Tim-Agro/S2-PDD.
Why it matches plant phenotyping methodsUAV RGB・ハイパースペクトル画像からジャガイモ個体の病害状態を推定する融合モデルを開発・比較検証し、公開データセットも構築しており、病害表現型の取得・抽出手法が中心である。
abstractThis study presents a spatial–spectral fusion deep learning framework (S2-PDD) that fuses uncrewed aerial vehicles (UAV)-based RGB imagery and hyperspectral-derived feature representations for plant-level detection of potato diseases (blackleg and Potato virus Y (PVY))
Abstract Grapevine leaves have dorsiventral anatomy with distinct adaxial (upper) and abaxial (lower) surfaces. Although morphological descriptor lists and ampelographic literature provide information on both the upper and lower side characteristics, in practice, the color traits of the upper side have become the focus of scientific publications. This study introduces the practical application of the recently developed LeafLaminaMap software and the use of trichromatic color indices in grapevine characterization. We aimed to compare colorimetric information on the adaxial and abaxial leaf surfaces as well as to explore the potential of machine learning models in classification. Five statistical descriptors (mean, standard deviation, contrast, energy, and entropy) were calculated for 25 RGB-based color indices on both leaf surfaces of 120 samples collected from four grapevine cultivars (‘Chardonnay’, ‘Pinot noir’, ‘Sauvignon blanc’, and ‘Syrah’). Data was subjected to multivariate statistical analysis and machine learning classifiers. Results showed that the abaxial leaf surface had stronger cultivar-specific color signatures, supporting its suitability for cultivar discrimination. These findings suggest that RGB-based analysis of both adaxial and abaxial leaf surfaces has potential for grapevine cultivar discrimination, offering a new perspective for cost-efficient plant phenotyping.
Why it matches plant phenotyping methodsRGB画像解析とLeafLaminaMap、色指数、機械学習を用いた葉面形質抽出・品種識別が研究の中心であり、植物フェノタイピング手法の実質的応用に該当する。
abstractThis study introduces the practical application of the recently developed LeafLaminaMap software and the use of trichromatic color indices in grapevine characterization.
Maize ear geometry (length, width, curvature, and volume) is closely tied to yield and grain-filling outcomes, but existing high-throughput phenotyping pipelines remain constrained by the cost, labor, and specialized hardware they require. We developed and validated a low-cost pipeline that reconstructs a watertight 3-D mesh of a maize ear from a single 20-second video captured with a consumer-grade DSLR on a motorized turntable under uniform LED illumination. Camera poses from a multi-seed COLMAP procedure initialize a Neural Radiance Field (NeRF), and a cylindrical holder of known diameter, visible in every frame, provides automatic metric scaling with downstream geometric quality control. Applied to 300 ears spanning a diverse maize inbred panel, 250 (83.3%) passed automated processing and quality control. Skeleton length agreed with manual caliper measurements across all 250 ears (R^2 = 0.964, RMSE = 4.68 mm), and convex-hull volume agreed with water-displacement volume on a 15-ear subset spanning the full size range (R^2 = 0.982, RMSE = 5.26 mL). Residual length error grew with ear curvature, whereas bounding-box height, which records the same straight-line chord as calipers, showed no such trend; the discrepancy therefore originates in the measurement definition, since calipers record the chord while skeleton length traces the geodesic arc. The capture hardware costs approximately 607 USD, and operator involvement fell from roughly five minutes to one minute per ear, with all downstream processing running unattended. The platform provides a foundation for breeding-scale 3-D ear phenotyping.
Why it matches plant phenotyping methodsトウモロコシ雌穂の3D形態形質を抽出する低コスト画像計測パイプラインを開発し、手動測定および体積測定で技術検証しているため、方法が研究の中心である。
abstractWe developed and validated a low-cost pipeline that reconstructs a watertight 3-D mesh of a maize ear
Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · checked 14 Sept 2026
Why it matches plant phenotyping methods画像から種子の形態・色形質を抽出し、生化学形質と統合したマルチモーダル表現型解析・機械学習分類法の開発と比較評価が研究の中心であるため。
abstractThis study developed an explainable machine learning framework integrating morphological, color, and biochemical characteristics for classifying chia ( Salvia hispanica L.) genotypes.
Goal. To substantiate the methodical approaches to analyzing the state of soybean crops based on the results of aerial photography with UAV by comparing manual vectorization, controlled classification according to the algorithm of maximum similarity, and uncontrolled classification of K-Means, as well as to determine the feasibility of their combination with expert visual interpretation to assess the spatial structure of the vegetation cover. Methods. Aerial photography of the test proving ground was performed with the help of the unmanned aerial vehicle DJI Phantom 4 Advanced with the subsequent photogrammetric study of materials and the formation of a highly detailed orthophotoplane. In the QGIS environment, visual decryption and manual vectorization of the main objects of the agrolandscape were carried out with the creation of polygonal layers. For automated mapping, methods of controlled classification according to the algorithm of maximum similarity and uncontrolled classification based on the K-Means algorithm were used. The accuracy of the results was evaluated by comparing the data of automated classifications with the data of manual digitization, which was used as a reference (control) method. On the basis of the results obtained, empirical data were summarized to justify practical recommendations for the application of the studied approaches. Results. The study was conducted on the territory of the research farm of the Separate subdivision of the National University of Life and Environmental Sciences of Ukraine «Berezhany Agrotechnical Institute» (vil. Pavliv, Ternopil district, Ternopil oblast) (49.452057°N; 24,805818°E) in may – august 2025. The obtained cartographic materials made it possible to quantify the areas of the main objects of the agro-landscape and identify problem areas with sparse shoots. Methods of controlled classification showed greater compliance with the digitization data (average deviation — 6.7%) compared to uncontrolled (14.7%), which were effective from the point of view of preliminary assessment of spectrally homogeneous sections, but did not provide an accurate division of shoots by density. Analysis of the spatial structure of coverage made it possible to plan local agrotechnical measures and assess the potential yield. Conclusions. Methodical approaches to analyzing the state of soybean crops based on the results of aerial photography with a UAV equipped with RGB cameras are promising and economically feasible. Automated classification methods are effective for highlighting hard and contrasting objects and small-contoured areas, while a detailed assessment of the structure of the vegetation cover is advisable to carry out using a controlled classification in combination with expert visual interpretation.
Why it matches plant phenotyping methodsUAV画像を用いて大豆作物の植生被覆構造や疎な出芽域を抽出し、複数の分類法を手動ベクトル化と比較・精度評価している。植物状態の取得手法自体が研究の中心である。
abstractTo substantiate the methodical approaches to analyzing the state of soybean crops based on the results of aerial photography with UAV by comparing manual vectorization, controlled classification according to the algorithm of maximum similarity, and uncontrolled classification of K-Means
Cauliflower emergence rate and seedling growth are key indicators of field conditions and varietal potential. Traditional manual surveys are unsuitable for continuous monitoring across multiple varieties. This study integrates UAV RGB imagery with the DualSlim-YOLO model to estimate cauliflower emergence rates and monitor seedling growth. Built on YOLOv11, the model incorporates a lightweight feature extraction structure and an optimized detection-scale configuration. It reduces computational complexity while maintaining detection accuracy, thereby improving the efficiency of cauliflower seedling detection. DualSlim-YOLO achieved P, R, F1-score, mAP@0.5, and mAP@0.5:0.95 of 95.35%, 96.75%, 96.05%, 98.55%, and 86.65%, respectively. The number of parameters was reduced by 38.61%, while the inference speed increased by 22.16%, demonstrating good lightweight performance. Based on this model, UAV images of 171 cauliflower varieties acquired at 7, 21, and 28 d after transplanting were used for seedling detection and emergence rate estimation. In addition, 18 time-series seedling phenotypic traits were extracted, enabling a comprehensive quantitative evaluation of emergence dynamics and early-growth performance across multiple cauliflower varieties. This method effectively screens cauliflower varieties for high emergence rates, rapid emergence, and excellent seedling growth performance. It provides technical support for high-throughput, nondestructive seedling phenotyping and early germplasm screening under field conditions.
Why it matches plant phenotyping methodsUAV画像と軽量YOLOモデルを用いて、カリフラワー苗の検出、出芽率推定、18種類の時系列表現型形質抽出を行う手法が中心であり、モデル性能も検証している。
abstractThis study integrates UAV RGB imagery with the DualSlim-YOLO model to estimate cauliflower emergence rates and monitor seedling growth.
Background: Faba bean is an important legume crop valued for its nutritional and soil-enriching benefits, yet its productivity is severely affected by foliar diseases. Automated image-based detection using deep learning provides a rapid and reliable approach for early disease identification and improved crop management. Methods: This study developed a machine learning-based automated framework for multi-class classification of Faba bean leaf diseases using transfer learning with the VGG16 convolutional neural network. A dataset of 8,021 RGB images collected under natural field conditions was used, comprising four classes: healthy, rust, gall and chocolate spot. Images were resized to 224 × 224 pixels and normalized prior to training. The pretrained convolutional layers of VGG16 were frozen and a custom classification head with global average pooling and dropout regularization was added. Model performance was evaluated using classification metrics. Result: The proposed model achieved an overall classification accuracy of 92.34% and a macro-averaged F1-score of 0.9227 on the test dataset. Strong classification performance was observed across all disease categories, with particularly high predictive accuracy for healthy and rust classes. The findings demonstrate the effectiveness of transfer learning for plant disease detection and highlight its potential for scalable, automated crop health monitoring in precision agriculture.
Why it matches plant phenotyping methods植物葉画像から病害状態を自動推定する深層学習手法の開発・性能評価が中心であり、植物フェノタイピング手法に該当する。
abstractThis study developed a machine learning-based automated framework for multi-class classification of Faba bean leaf diseases using transfer learning with the VGG16 convolutional neural network.
Plant diseases are still posing a challenge to the productivity, quality of crops, and food security, especially in locations where field diagnosis is based on manual visual inspec-tion. This paper assesses deep learning network-based automated classification of plant leaf diseases on public RGB leaf-image datasets, such as the Kaggle New Plant Diseases Dataset (Augmented) and PlantVillage images. They investigated four archi-tectures: EfficientNetV2B0, ResNet152V2, DenseNet201, and one hybrid Vision Trans-former (ViT)-based model. The steps of the experiment involved loading the dataset, exploratory analysis, preprocessing, resizing, normalizing, augmentation, transfer learning, independent model training, and evaluation metrics such as accuracy, preci-sion, recall, F1-score, training curves, testing results, and confusion matrices. The hy-brid ViT-based model was reported to have the best accuracy of 99.5%. On the smaller seven class subset of PlantVillage, EfficientNetV2B0 scored 98.11%. On the 38-class dataset, DenseNet201 improved test accuracy (97.34) and validation classification ac-curacy (around 98). ResNet152V2 scored 97.01 on the 38-class test set. The results demonstrate that CNN and transformer-based models can help to recognize plant diseases accurately whereas hybrid attention-based structures provide a promising path to enhance fine-grained classification. Since the model notebooks had varying class settings and splits, the comparison is seen as a model-structured assessment as opposed to a precisely identical benchmark across all architectures.
Why it matches plant phenotyping methods植物葉画像から病害状態を分類する深層学習手法を複数モデルで比較・評価しており、病害表現型の取得・抽出と技術検証が研究の中心である。
abstractThis paper assesses deep learning network-based automated classification of plant leaf diseases on public RGB leaf-image datasets
Reproduction assets foundThe paper's phenotyping inputs are two public plant leaf-image datasets explicitly named in the Data Availability statement: the Kaggle New Plant Diseases Dataset (Augmented) and the PlantVillage dataset, both with public URLs. No author code, models, or supplementary materials are deposited (supplementary materials: 'Dataset · publicy available. Plant leaf images were
obtained from the Kaggle New Plant Diseases Dataset (Augmented) and PlantVillage
datasets. The datasets contain publicly accessible RGB images of healthy and diseased
plant leaves used for supervised image classification research.
DATASET SOURCES
Kaggle New Plant Diseases Dataset (Augmented):
https://www.kaggle.com/datasets/vipoooool/new-plant-diseases-dataset
PlantVillage Dataset: https://plantvillage.psu.edu/All processed data, experimental configurations, and model implementation details are
described within the manuscript. Additional materials may be made available from the
corresponding author upon reasonable request.
ACKNOWLEDGMENTS
The author acknowOpen asset ↗Kaggle · new-plant-diseases-datasetpdf-raw-page:24 lines:1-23Dataset · publicgmented) and PlantVillage
datasets. The datasets contain publicly accessible RGB images of healthy and diseased
plant leaves used for supervised image classification research.
DATASET SOURCES
Kaggle New Plant Diseases Dataset (Augmented):
https://www.kaggle.com/datasets/vipoooool/new-plant-diseases-dataset
PlantVillage Dataset: https://plantvillage.psu.edu/All processed data, experimental configurations, and model implementation details are
described within the manuscript. Additional materials may be made available from the
corresponding author upon reasonable request.
ACKNOWLEDGMENTS
The author acknowledges Istanbul Aydin University for academic support and research
guidance duringOpen asset ↗PlantVillagepdf-raw-page:24 lines:1-23Code / dataset availability confirmedCrossref · checked 15 Sept 2026
Precision agriculture is becoming more and more of a challenge that requires the use of intelligent systems that are able to predict stress and prevent yield loss before it is too late. Traditional methods of agricultural surveillance are predominantly reactive with irrigation demands being based on thresholds or individual yield forecasts models that do not represent the intricate spatio-temporal interactions that exist between crop physiology, soil status, and environmental stresses. Besides, the majority of the current practices do not have an autonomous decision-making approach to preventive intervention which leads to inefficient use of water and slows down the response to stress. This paper suggests a cognitive UAV-assisted agro-surveillance system to predict yield vulnerability caused by crop stress and optimize adaptive irrigation with the help of spatio-temporal deep and reinforcement learning. The framework combines UAV-obtained RGB and multispectral and thermal imagery with measurements of soil sensors and meteorological data obtained with the Crop Health and Environmental Stress Dataset. A new GeoSpatio-TRiNet model is used to acquire long-range spatial relationship, time stress development, and diffusion of stresses across agricultural regions. The model predicts the vulnerability trajectories of the stress instead of the direct yield regression, and this allows early detection of yield risk. Such predictions serve to generate a cognitive environmental state of a Soft ActorCritic (SAC) reinforcement learning agent that autonomously computes zone-based irrigation behaviors to reduce the recurrence of stress at the minimum water usage cost. As shown by the results of the experiment, the proposed framework has a stress forecasting accuracy of 96.3% and performs much better than the traditional machine learning, CNN-based, and transformer-based baselines. The system also decreases the predicted yield vulnerability by 46.6 and enhances water-use efficiency by 41.1 as compared to irrigation strategies based on rules. The results confirm the usefulness of spatio-temporal intelligence with predictive control in terms of effectiveness, and the proposed framework is a scalable and sustainable solution to precision agriculture of the next generation.
Why it matches plant phenotyping methodsUAV画像とセンサーデータから作物ストレスの時系列状態および収量脆弱性を推定する計算・センシング手法が研究の中心であり、灌漑制御への応用も技術評価の一部として記述されている。
abstractThe framework combines UAV-obtained RGB and multispectral and thermal imagery with measurements of soil sensors and meteorological data
Reproduction assets foundThe paper uses the public Kaggle Crop Health and Environmental Stress Dataset (UAV RGB/multispectral/thermal imagery plus soil/weather measurements and stress labels) as its phenotyping data source, and the authors provide an explicit public GitHub repository for the analysis code.Dataset · publicThe current research is based on the Crop Health and Environmental Stress Dataset, which is a publicly available
dataset on Kaggle, specially created to help perform a spatio-temporal analysis of crop health in response to changing
environmental and water-stress factors [26].Open asset ↗pdf-raw-page:10 lines:1-62Code · publicturn: Final zone-wise stress predictions 𝐶
𝑡
𝑧, Yield vulnerability trajectories 𝑉𝑡
𝑧, Optimal adaptive irrigation policy
𝜋∗
End Algorithm
Code availability:
The data used to support the findings of this study are included in the article.
Code availability:
The code used in this research work is available in the following link.
https://github.com/replyvenugopal/Cognitive-UAV-Driven-Agro-Surveillance
4. Result and Discussion
The architectural agro-surveillance solution, which is proposed to be executed by UAVs, is executed through a
modular and scalable software framework to guarantee reproducibility and extensibility. The experiments are all
performed in Python as a main programming languageOpen asset ↗github.com/replyvenugopal/Cognitive-UAV-Driven-Agro-Surveillancepdf-raw-page:24 lines:1-55Plant phenotyping relevance match · UnverifiedCrossref · checked 5 Sept 2026
Chlorophyll fluorescence provides sensitive information on plant photochemical responses, but its measurement requirements can limit high-throughput application. This study investigated whether RGB imagery could approximate chlorophyll-fluorescence-derived photochemical status across garden plant species during progressive soil drying. A Photochemical Status Index (PSI) was constructed by principal component analysis from five highly correlated JIP-test energy-flux variables (RC/CS, ABS/CS, TRo/CS, ET2o/CS, and RE1o/CS). The dataset comprised 50 aggregated species-by-soil-moisture-stage observations representing ten species and five sequential soil-moisture stages. Eleven RGB-derived variables were evaluated, and a partial least-squares regression model was assessed using nested leave-one-species-out validation, with all data-dependent procedures repeated within each outer training fold. PC1 explained 96.5% of the shared variation among the fluorescence-derived fluxes. The predictors g, GLI, ExG, ExGR, and CIVE were retained in all ten outer folds. The final model yielded a pooled out-of-fold R2 of 0.469, an RMSE of 1.585, and an MAE of 1.183. However, species-specific R2 ranged from −0.179 to 0.959, and a calibration slope of 0.509 indicated prediction-range compression. These findings provide proof-of-concept evidence of moderate RGB-based approximation of fluorescence-derived photochemical status, but inconsistent species transferability and the common soil-moisture/time gradient require external validation before practical deployment.
Why it matches plant phenotyping methodsRGB画像から蛍光由来の植物光化学状態を推定する手法を構築し、種間交差検証で性能評価しており、植物フェノタイピング手法が中心である。
abstractThis study investigated whether RGB imagery could approximate chlorophyll-fluorescence-derived photochemical status across garden plant species during progressive soil drying.
High-resolution monitoring of forest structure and productivity is essential for effective natural resource management. However, monitoring approaches such as field-based forest inventories or extensive lidar campaigns are costly, time-intensive, and spatially limited. Therefore, inexpensive and accessible methods are needed. SatCHM (Satellite Canopy Height Model) was developed to be an accessible and open-source tool for researchers, allowing for site-specific and temporally flexible predictions of canopy height with limited computational resources. SatCHM requires four inputs: panchromatic satellite imagery, solar and sensor angle metadata of satellite imagery, digital elevation models (DEMs), and lidar-produced CHMs for an area of interest. After SatCHM pre-processes inputs, data is loaded into a collection of convolutional neural networks (CNNs) for image-to-image regression. This ensemble cooperates to yield high-resolution predictions (up to 0.5-meter) of three-dimensional tree structure with discernible tree crowns across a broader defined area of interest. After calculating the mean absolute error for each prediction output, the median of these mean absolute errors was 6.06 meters.
Why it matches plant phenotyping methods森林キャノピー高と樹冠構造という植物形質を衛星画像等から推定するオープンソース手法を開発し、CNNによる推定と誤差評価まで行っており、植物フェノタイピング手法が研究の中心である。
abstractSatCHM (Satellite Canopy Height Model) was developed to be an accessible and open-source tool for researchers, allowing for site-specific and temporally flexible predictions of canopy height with limited computational resources.
ABSTRACT Leaf shape is a fundamental trait of plant ecological strategies, influencing biotic interactions and ecosystem functioning. However, established quantitative metrics fail to capture subtle variations and irregularities, require user-based reference points or are challenging to compare among taxa with broadly different leaf shapes. In addition, established metrics typically conflate (aggregate) leaf edge complexity and macro-shape complexity, despite their independent functional significance and genetic foundations. Here, we introduce an entropy-based framework to quantify two new complexity metrics: edge complexity and macro-shape complexity. Based on three case studies, we show that these metrics outperform aggregate metrics in predicting Quercus robur chemical traits, provide more intuitive interspecific classifications, and strongly align with human perception. In addition, edge and macro-shape complexity show high complementarity, while aggregate metrics are highly redundant and typically strongly related to leaf area. Emerging as the strongest predictor of leaf chemistry and key visual cue for complexity as perceived by humans, the effects of edge complexity highlight the under-appreciated functional significance of leaf margins. Our framework and the proposed entropy-based complexity metrics thus promise to help unlock the potential of growing digital image archives of leaves, including images from herbaria and fossils, and are technically readily applicable to shapes of algae, bacteria, pollen, and beyond. The accompanying package ShapeComplexity enables the broad application of entropy-based metrics, providing a powerful tool to explore how the shape of organisms and biological structures influences ecological strategies, biotic interactions, and ecosystem functioning while tracking spatial and temporal variation.
Why it matches plant phenotyping methods葉の画像からエッジ複雑性とマクロ形状複雑性を定量化する新規指標とソフトウェアを開発しており、植物形質抽出法が研究の中心である。
abstractHere, we introduce an entropy-based framework to quantify two new complexity metrics: edge complexity and macro-shape complexity.
Reproduction assets foundThe paper's authors publicly release their ShapeComplexity analysis code (Rust) on GitHub, used to compute the paper's leaf edge- and macro-shape complexity metrics. Supplementary data/analysis code are on Dryad, but that URL is not in the allowed list. RMBG is a generic third-party background-removal model, not a phenCode · publicThe complete, open-source Rust-code (The Rust Team, 2025 ) is publicly available on GitHub ( https://github.com/Thornbach/ShapeComplexity ), ensuring transparency and reproducibilityOpen asset ↗Thornbach/ShapeComplexitylines:86-94Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · checked 14 Sept 2026
Abstract Ascochyta blight is a widely occurring chickpea fungal disease that can cause severe yield loss. Breeding for crop resistance benefits from high‐throughput evaluation of plant–pathogen interactions in genotypes which can serve as sources of resistance. Current practice for the evaluation is human visual scoring of disease symptoms, which is limited in throughput and precision. Here, we developed open‐source sensor‐based phenotyping methods using red, green, blue (RGB) and multispectral imaging to measure resistance components and predict disease severity classes in chickpea and wild relatives grown outdoors over three seasons. Pots were imaged at multiple time points with a ground‐based platform, providing 86,792 RGB and 8199 multispectral images. Lesion count was estimated with YOLOv5 (You Only Look Once version 5) object detection (F1 score = 0.27–0.30), fractional green canopy cover was estimated from RGB images, and vegetation indices were extracted from multispectral images. A model trained on growth rates of fractional green canopy cover normalized to control genotypes could predict disease severity classes with an accuracy of 65% –81 % ( 0.43–0.59) on unseen data from three different seasons. The developed methods provide a pathway to predict visual disease severity scores and support the breeding of crops for disease resistance. They may also be used to characterize disease progression, to find underlying resistance mechanisms, and for early disease detection.
Why it matches plant phenotyping methodsRGB・マルチスペクトル画像と地上センサープラットフォームを開発し、病斑数、緑色キャノピー被覆率、病害重症度を推定・予測する手法が研究の中心であるため。
abstractHere, we developed open‐source sensor‐based phenotyping methods using red, green, blue (RGB) and multispectral imaging to measure resistance components and predict disease severity classes in chickpea and wild relatives grown outdoors over three seasons.
Abstract This study evaluated the potential of UAV-derived RGB spectral indices to predict fiber quality traits in cotton genotypes. Nineteen genotypes were assessed under a randomized complete block design, and RGB imagery acquired at full flowering was used to calculate GLI, NGRDI, SCId, and SI. Significant genetic variability and moderate-to-high heritability were observed for both fiber traits and spectral indices. GLI was positively associated with the Spinning Consistency Index, whereas NGRDI was associated with fiber length uniformity. Regression models showed moderate predictive ability (R² LOOCV between 32.4–35.4%; and accuracy between 0.57–0.60). GLI and NGRDI demonstrated potential as complementary tools for large-scale phenotyping and preliminary genotype selection, although they do not replace conventional fiber quality analyses. Further studies across additional developmental stages are needed to improve prediction accuracy.
Why it matches plant phenotyping methodsUAV由来RGB画像からスペクトル指標を算出し、ワタの繊維品質形質を予測する手法を評価しており、大規模フェノタイピングへの応用と予測性能の検証が中心である。
abstractThis study evaluated the potential of UAV-derived RGB spectral indices to predict fiber quality traits in cotton genotypes.
Northeast China's Geng rice (Oryza sativa subsp. japonica) dominates the high-value rice markets in China due to its superior eating quality. However, current evaluation methods rely on either labor-intensive, subjective sensory protocols or low-accuracy, calibration-heavy near-infrared spectroscopy (NIRS), constraining breeding for high eating quality and market development. Here, we report a vision-based deep learning framework combining multi-population fine-tuning with industrial vision-language model (VLM) pre-training for Geng rice eating quality prediction. Trained on natural and recombinant inbred (RI) population datasets, our optimal model (Model 4) showed high cross-population stability. It achieved R 2 values of 0.98, 0.57, and 0.61 in a natural population validation set (35 cultivars), an independent DA-RI population (201 lines), and a randomly collected set (30 Northeast and 28 Southern cultivars), respectively, consistently outperforming the widely used Satake STA1B analyzer. Furthermore, our approach enabled the mapping of a novel, robust quantitative trait locus, qIVOE7, for Geng rice eating quality on Chromosome 7. Further analysis suggested that Model 4 appears to rely on the Hue dimension of the HSV color space for its predictions. This framework provides a high-accuracy prediction model and an easy-to-use tool for rice eating quality evaluation, accelerating high-quality rice breeding as well as the development of the high-quality rice market.
Why it matches plant phenotyping methodsコメの食味という植物（種子）形質を画像ベースの深層学習で推定する手法を開発し、複数集団で検証・既存分析器と比較しており、フェノタイピング手法が研究の中心である。
abstractwe report a vision-based deep learning framework combining multi-population fine-tuning with industrial vision-language model (VLM) pre-training for Geng rice eating quality prediction.
Accurate and non-destructive assessment of drought stress is important for improving lettuce production and supporting timely crop management. This study presents a detection-guided deep learning framework for plant-level drought-stress assessment in hydroponically grown lettuce using bird’s-eye-view RGB images. The study further investigates whether canopy segmentation can improve classification performance by reducing irrelevant background information. The framework was evaluated using 2190 images collected across three independent cultivation cycles in which drought stress was induced by isolating the plant root zones from the nutrient solution. In the first stage, YOLO-based object detection was used to localize individual plants, with YOLO26m achieving the highest detection performance of 99.4% mAP@0.5. The detected regions were subsequently used as spatial prompts for zero-shot canopy segmentation using the Segment Anything Model (SAM), with SAM ViT-B achieving a mean IoU of 0.9864. Six convolutional, transformer-based, and hybrid classification architectures were then evaluated independently using YOLO-cropped and SAM-segmented plant images. Segmented inputs consistently improved classification performance, with MaxViT-S achieving the highest binary test accuracy of 96.3%. The framework further distinguished time-defined pre-stress, early-stress, and late-stress periods with an accuracy of 92.4%. Plant-level generalization was further assessed using six-fold leave-one-plant-out cross-validation, resulting in a mean test accuracy of 90.25 ± 1.78% on unseen plants. These findings demonstrate that RGB-based plant-level analysis can support non-destructive drought-stress assessment and that canopy segmentation improves classification by reducing background influence.
Why it matches plant phenotyping methodsRGB画像からレタス個体の乾燥ストレス状態を推定する検出・セグメンテーション・分類フレームワークを開発し、複数サイクル、未見個体、性能指標で検証しており、表現型取得手法が中心である。
abstractThis study presents a detection-guided deep learning framework for plant-level drought-stress assessment in hydroponically grown lettuce using bird’s-eye-view RGB images.
Accurate quantification of plant disease severity is essential for evaluating host-pathogen interactions and assessing the effectiveness of disease management strategies. Traditional visual scoring methods and manual estimation of infected tissue are widely used but are often subjective and prone to observer bias. Digital image analysis offers an objective alternative by enabling automated identification and quantification of symptomatic plant tissues based on color and spatial characteristics. Here, we present a MATLAB-based image processing protocol for differentiating diseased and healthy plant tissue from digital leaf images. The workflow involves acquisition of standardized leaf images, conversion of RGB images into hue-saturation-value (HSV) color space, segmentation of diseased tissue using defined HSV thresholds, refinement of the segmented mask through morphological operations, and extraction of the whole leaf area. The protocol then calculates the diseased area and total leaf area in pixels and computes the percentage of infected tissue. The method uses MATLAB together with the Image Processing Toolbox and can be implemented using simple scripts. This protocol enables rapid and reproducible quantification of disease severity in plant leaves exhibiting visually distinct symptoms such as necrotic lesions or blight patches. By minimizing observer bias and providing quantitative measurements of infected area, the protocol offers a practical and reproducible approach for plant disease phenotyping and evaluation of disease management strategies across diverse plant-pathogen systems where diseased tissues can be clearly distinguished from healthy tissues under reasonably controlled imaging conditions. Key features • A reproducible MATLAB-based workflow for separating diseased and healthy plant tissue using color-space segmentation. • Applicable to plant diseases where symptomatic tissue contrasts clearly with healthy tissue (necrosis, blight lesions, rot patches). • Requires digital leaf images, MATLAB, and the MATLAB Image Processing Toolbox for image processing and disease quantification. • Enables rapid calculation of diseased leaf area and disease severity using automated pixel-based quantification.
Why it matches plant phenotyping methods植物病斑を画像から分割・定量し、感染面積と病害重症度を算出するMATLAB画像解析プロトコルが研究の中心であり、植物病害表現型の取得・抽出手法に該当する。
abstractHere, we present a MATLAB-based image processing protocol for differentiating diseased and healthy plant tissue from digital leaf images.
Reproduction assets foundThe protocol explicitly deposits its authors' MATLAB image-processing workflow (HSV segmentation, mask refinement, pixel-based disease quantification) in a public GitHub repository with README instructions and example images.Code · publicGitHub repository containing the MATLAB source code, README file with installation and execution instructions, and representative example image(s): https://github.com/pankajborahmajuli-source/Leaf-Disease-Detection-MATLAB-Code/blob/main/README.mdOpen asset ↗Leaf-Disease-Detection-MATLAB-Codehtml-lines:112-148Plant phenotyping relevance match · UnverifiedCrossref · checked 15 Sept 2026
Food safety globally is threatened by crop disease, which creates a major obstacle to yield losses, so there is an urgent need for rapid, precise, and large-scale diagnostic methods for all the global risks crops are exposed to from disease. While imaging sensors, as well as Artificial Intelligence (AI), have made great strides in recognising plant disease, most literature does not have a comprehensive analysis that combines methods, technology, and implementation. Therefore, a systematic literature review follows PRISMA methods; we review 61 excellent studies published within the last five years that outline the advancement of imaging modalities (Red, Green, Blue (RGB), multispectral/ hyperspectral, thermal), deep learning architectures, augmentation of data, explanation methods and IoT (Internet of Things)-edge-cloud for managing intelligent agriculture. These modern AI-based systems (AI systems) have consistently produced accurate results above 98%. However, there are problems with the generalisability (across hybrid plant species), robustness (when exposed to environmental stresses), and interpretability of the results presented to consumers. This review represents the first compilation of using imaging sensors, artificial intelligence models, Internet of Things architecture (IoT-edge), and robotics into one comprehensive framework for the detection of plant disease in the next generation. In addition, this review suggests future research directions, including lightweight edge-deployable models, multimodal sensor fusion, interpretable AI, larger validated datasets, and autonomous robotic systems for scalable and sustainable smart agriculture.
Why it matches plant phenotyping methods植物病害の画像・センサーによる検出手法を体系的にレビューしており、病害状態のフェノタイピング手法が中心です。
titleComprehensive Review of Plant Disease Detection: Advancements in Imaging Sensors, AI Techniques, and Future Directions in Smart Agriculture
Localized soil-moisture deficits, that is, irregular sub-field patches where crops experience water stress well before visible wilting, are a leading cause of yield variability in row-crop agriculture. These zones are difficult to detect at the spatial resolution and revisit frequency required for timely irrigation response. This paper presents a reinforcement-learningguided autonomous quadrotor unmanned aerial vehicle (UAV) platform that fuses onboard Visual Simultaneous Localization and Mapping (Visual SLAM) with a pushbroom hyperspectral imaging payload to construct georeferenced, canopy-registered maps of a Crop Water-Stress Index (CWSI) in near real time. Rather than flying a fixed lawnmower survey, the platform is guided by an adaptive-sampling policy trained with Proximal Policy Optimization (PPO) that reallocates flight time and sensor dwell toward regions of emerging water stress as evidence accumulates mid-flight. We present the complete engineering pipeline: airframe and sensor design, a keyframe-based Visual SLAM front and back end that provides centimeter-scale geolocation without continuous reliance on Real-Time Kinematic (RTK) GNSS lock, a hyperspectral preprocessing and spectralindex chain (NDVI, NDRE, NDWI/NDMI) used to derive CWSI through a learned regression, the partially observable Markov Decision Process (POMDP) formulation and reward shaping used to train the sampling policy, and the fused system architecture tying these subsystems together. In simulated field trials over a 0.8-hectare test plot, the reinforcement-learning-guided policy achieved a 92% water-stress-zone detection rate versus 61% for a fixed-grid baseline, while reducing mission flight time by approximately 32%. We further report an ablation study isolating the contribution of SLAM-derived canopy structure to CWSI accuracy, a sensitivity analysis across field complexity, and a full error budget for the fused pipeline. We close with a discussion of validation limitations, broader scientific and agricultural impact, and a roadmap toward multi-UAV fleet deployment for whole-farm monitoring
Why it matches plant phenotyping methodsUAV、Visual SLAM、ハイパースペクトル画像、機械学習を統合し、作物の水ストレス状態を推定・地図化する技術パイプラインを開発・評価しており、植物表現型取得が中心である。
abstractThis paper presents a reinforcement-learningguided autonomous quadrotor unmanned aerial vehicle (UAV) platform that fuses onboard Visual Simultaneous Localization and Mapping (Visual SLAM) with a pushbroom hyperspectral imaging payload to construct georeferenced, canopy-registered maps of a Crop Water-Stress Index (CWSI) in near real time.
Soil salinity is a major constraint on rice ( Oryza sativa L.) production, particularly during the yield-determining reproductive stage. Utilizing a high-throughput RGB platform, we non-destructively phenotyped a diverse panel of 294 rice accessions over two consecutive years. By extracting 60 dynamic image-based traits (i-traits) reflecting canopy architecture and stay-green capacity, and four seed-setting rate-related traits, our genome-wide association study (GWAS) identified 95 significant loci, 35.8% of which precisely co-localized with previously reported QTLs. We further prioritized OsSLT1 ( LOC_Os01g05790 ) as a candidate gene at a reproducible suggestive locus associated with leaf-rolling-related image traits. Transgenic evaluations confirmed that it acts as a positive regulator of salt tolerance at the seedling stage. OsSLT1 was mainly detected in the nucleus, and no significant changes in Na + or K + accumulation were observed in flag leaves under the tested salt-stress condition, suggesting that OsSLT1 may regulate salt tolerance through mechanisms beyond classical shoot ion accumulation. Natural variations in the OsSLT1 promoter were associated with transcriptional divergence. The Hap2 promoter haplotype showed significantly higher stress-induced transcriptional activity. Hap2 was rare in modern indica accessions, suggesting that it may represent a potentially useful genetic resource for future salt-tolerance improvement. Supplementary information The online version contains supplementary material available at 10.1007/s11032-026-01704-2.
Why it matches plant phenotyping methodsRGB高スループット基盤による非破壊画像計測と、60種の動的画像形質の抽出が研究の主要なデータ取得・解析手法として明示されているため、GWAS中心の応用研究でも植物フェノタイピング手法の実質的応用に該当する。
abstractUtilizing a high-throughput RGB platform, we non-destructively phenotyped a diverse panel of 294 rice accessions over two consecutive years.
The edible yield and commercial value of durian are strongly determined by the number of fully developed internal locules, yet their assessment still relies largely on subjective manual inspection or costly destructive analysis. Existing non-destructive approaches, including tapping-based evaluation and X-ray imaging, are either insufficiently standardized or impractical for high-throughput phenotyping. This study proposes the Rotation-Synchronized Structural Locule Evaluator (RS-SLE). The framework integrates convolutional spatial learning with deterministic signal processing to estimate internal locule number from rotational RGB imaging. The proposed system learns frame-wise spatial morphology using convolutional heatmap regression and performs temporal inference through deterministic kinematic signal processing over a mechanically bounded 360° rotation, thereby avoiding reliance on data-intensive recurrent video models. Specifically, sequential 2D locule-probability heatmaps are transformed into a synchronized 1D morphological energy signal, which is subsequently refined using gradient-based integration, Tikhonov regularization, and adaptive morphological thresholding to recover cycle-consistent structural peaks corresponding to fertile locules. This design provides a transparent and computationally efficient alternative to learned temporal memory while maintaining robustness to viewpoint variation, partial occlusion, and high-frequency spine noise. Evaluated on 260 fruits of the ‘Monthong’ cultivar, the proposed framework achieved a mean absolute error of 0.289 locules and a 71.1% exact-match rate on independent rotational videos. When combined with morphological correlation analysis, the framework attained a 100% tolerance accuracy (±1 locule) on the test set with destructive ground-truth measurements. These results demonstrate that dynamic external morphology can serve as a reliable optical proxy for internal locule development. Crucially, the final locule count is derived entirely from the CNN-generated one-dimensional morphological signal. This demonstrates the necessity of localized structural analysis over simple macroscopic shape indices. The findings highlight the value of explainable, kinematics-informed vision systems for practical precision agriculture.
Why it matches plant phenotyping methods回転RGB画像からドリアン内部のlocule数という植物器官状態を推定する画像ベース表現型計測法を開発し、独立動画と破壊的グラウンドトゥルースで検証しているため、方法が研究の中心である。
abstractThis study proposes the Rotation-Synchronized Structural Locule Evaluator (RS-SLE).
Quantitative flowering phenotypes are needed to support breeding and harvest management in Hypericum perforatum L. (St. John’s wort), but manual flower assessment is slow and difficult to standardize under field conditions. UAV-Hyp is a multi-temporal UAV RGB dataset containing 12,653 high-resolution images acquired at 26 measurement dates across the complete flowering period of 15 H. perforatum accessions. The images represent variable illumination, soil moisture, weed pressure, and developmental stages. The dataset provides 59,163 plant bounding boxes and 107,054 flower bounding boxes. As an application example, cascaded YOLOv8 plant and flower detectors achieved mAP@0.50:0.95 values of 0.977 and 0.950, respectively. UAV-Hyp supports scalable flower quantification and the development of time-series phenotyping methods for genotype comparison and quality-oriented medicinal-plant breeding.
Why it matches plant phenotyping methods植物の開花形質を定量化するUAV画像データセットを提供し、検出性能も評価しているため、フェノタイピング用データセット・解析手法が中心です。
abstractUAV-Hyp is a multi-temporal UAV RGB dataset containing 12,653 high-resolution images acquired at 26 measurement dates across the complete flowering period of 15 H. perforatum accessions.
Calibrating the mutagenic dose is the first practical step of any radiation mutation-breeding programme, and it is usually summarised by the median lethal dose (LD50) or the median growth-reduction dose (GR50). We asked whether an accessible, image-based phenotyping pipeline can quantify the early radiation response of cowpea (Vigna unguiculata L. Walp.) seedlings finely enough to estimate GR50 and to rank organ- and pigment-level sensitivities. Seeds of the traditional Paraguayan landrace kumandá pyta’i were exposed to Cobalt-60 gamma rays at 0, 100, 200, 300, 400, 500, 600, and 700 Gy, grown in a greenhouse, and photographed at the early seedling stage. A single calibrated photograph (5.1 px mm−1) of 83 seedlings was segmented in Fiji/ImageJ 1.54p and analysed with Python to extract morphometric traits (total, root, and shoot length, root:shoot ratio, tortuosity, and a two-dimensional biomass proxy) and colorimetric traits (CIE L*a*b*, a normalised greenness index, and colour-class pixel fractions). Because the data departed from normality, dose effects were tested with Kruskal–Wallis, Spearman rank correlation, and Dunn post hoc tests, and GR50 was estimated by regression of each trait expressed as a percentage of the control. Total length, shoot length, and the biomass proxy declined significantly with dose (Spearman ρ = −0.40, −0.51, and −0.47; all p < 0.001), preceded by a low-dose stimulation at 100 Gy. Estimated GR50 values were ≈390 Gy for shoot length, ≈510 Gy for total length, and ≈550 Gy for the biomass proxy, within the range reported for other cowpea genotypes. Shoot elongation was more radiosensitive than root elongation, so the root:shoot ratio did not decline; tortuosity showed no dose response. Among pigment traits, the loss of greenness was the most robust signal (a* increased, ρ = +0.62, p = 5 × 10−10; green pixel fraction fell from 0.32 to near zero by 500 Gy). These results show that single-photograph phenotyping resolves a coherent, statistically supported dose response and yields a GR50 estimate usable for dose calibration. For kumandá pyta’i, doses of roughly 300–400 Gy (below GR50) are the most defensible starting window for mutation induction. The framework is reproducible and low-cost, but it is based on one greenhouse experiment and a single genotype, and should be validated across independent trials and cultivars.
Why it matches plant phenotyping methods画像取得・セグメンテーション・解析による形態および色彩形質の抽出を中心に、放射線応答とGR50を推定する低コスト画像ベース表現型解析法を提示しているため。
abstractA single calibrated photograph (5.1 px mm−1) of 83 seedlings was segmented in Fiji/ImageJ 1.54p and analysed with Python to extract morphometric traits
Drought stress severely limits foxtail millet yield and quality, yet current drought-resistance indices are exclusively yield-oriented and ignore grain-filling quality. Our two-year (2024–2025) experiments with 24–48 varieties revealed that yield and blighted grain rate (BGR) are partially decoupled (e.g., Zhangzagu 18: yield 2307 kg/ha, BGR 0.444; Zhonggu 19: yield 1622 kg/ha, BGR 0.280). We therefore constructed the Yield–Quality Synergy Index (YQSI = DYI − BGR), which penalizes varieties with poor grain filling. The YQSI tied for first place with DYI in comprehensive screening performance and achieved the highest inter-annual stability (Spearman ρ = 0.823, Jaccard = 0.438, composite score = 1.261). Sensitivity analysis confirmed robustness of the equal-weight formula across a 4-fold range of quality-penalty weights. Six strongly drought-resistant germplasms with balanced yield and quality were identified. Using UAV multimodal data (RGB, multispectral, and thermal infrared) acquired during grain filling, a Random Forest model predicted a YQSI with overall R2 = 0.819 and an F1 score of 0.933 for variety screening. Feature-importance analysis highlighted NDVI, WDRVI, and red-edge texture as key predictors. This study provides a quality-constrained drought-resistance evaluation framework and demonstrates the potential of UAV-based high-throughput phenotyping for foxtail millet breeding.
Why it matches plant phenotyping methodsUAVのRGB・マルチスペクトル・熱赤外データから干ばつ耐性指標を予測する高スループット表現型解析手法が研究の中心であり、モデル性能も評価している。
abstractUsing UAV multimodal data (RGB, multispectral, and thermal infrared) acquired during grain filling, a Random Forest model predicted a YQSI with overall R2 = 0.819 and an F1 score of 0.933 for variety screening.
Existing reviews on AI in tea production are either agriculture-generic or limited to isolated tasks. This review thoroughly compares vision technologies (RGB, hyperspectral, near-infrared, thermal, Light Detection and Ranging (LiDAR), Unmanned Aerial Vehicle (UAV)) and establishes a task-oriented algorithm selection framework for the tea industry. For small-sample or near-linear problems, traditional machine learning (ML) (support vector machine (SVM); partial least squares regression (PLSR)) remains effective. For unstructured field tasks, deep learning achieves superior performance: pest detection accuracy exceeds 97%, tea bud detection reaches 96.8% with RGB images, and hyperspectral imaging predicts nitrogen content with R 2 > 0.90 and tea polyphenols with R 2 up to 0.925. Algorithm choice further differentiates by task granularity: lightweight convolutional neural networks (CNNs) balance speed and accuracy for edge deployment at 16 fps; You Only Look Once (YOLO) series detectors enable real-time localization on mobile platforms at 93.1% accuracy, 24 ms per target. No single algorithm dominates all tea tasks; selection is a trade-off among accuracy, speed, data availability, and computational constraints. These findings outline a structured analysis of the challenges and pathways for transitioning computer vision (CV) from laboratory research toward field-deployable tools.
Why it matches plant phenotyping methods茶作物の画像センシング技術と解析アルゴリズムを体系的に比較し、害虫検出や窒素含量予測など植物状態・形質の推定方法を扱う方法論レビューである。
abstractThis review thoroughly compares vision technologies (RGB, hyperspectral, near-infrared, thermal, Light Detection and Ranging (LiDAR), Unmanned Aerial Vehicle (UAV)) and establishes a task-oriented algorithm selection framework for the tea industry.
This study investigates the automatic classification of crop phenological stages from low-altitude UAV RGB imagery. The dataset included 11,489 images of five crops: sunflower, rapeseed, soybean, wheat, and barley. The images were annotated using the Biologische Bundesanstalt, Bundessortenamt und Chemische Industrie (BBCH) scale, with labels corresponding to either single stages or stage ranges to reflect heterogeneous field conditions and transitional crop states. A pretrained ResNet18 model was adapted to the task using transfer learning. Training was conducted in two stages: first, the classification head was optimized while the backbone remained frozen; second, the entire network was fine-tuned. The model achieved strong internal test accuracy across all crops, with 100% test accuracy for rapeseed and barley, more than 99% for the remaining crops, and a mean accuracy of 99.73% under the studied survey conditions. The results also compare favorably with previously reported studies on UAV-based phenological classification. Overall, the findings support the potential of low-altitude UAV imagery and deep learning for localized phenological assessment of selected field zones in precision agriculture, while broader deployment requires validation across independent fields, seasons, regions, and survey conditions.
Why it matches plant phenotyping methodsUAV画像と深層学習により作物の生育（フェノロジー）段階を自動推定する手法が研究の中心であり、植物状態の取得・分類に直接関わる。
abstractThis study investigates the automatic classification of crop phenological stages from low-altitude UAV RGB imagery.
Accurate field detection of candidate tea shoots could support plantation monitoring, yield estimation, fresh-leaf assessment, and future selective-harvesting research. Wuyi rock-tea shoots are small and slender, have weak visual boundaries, and are easily confused with branches, petioles, and complex canopy backgrounds. Here, we developed YOLO11s-CSNG for candidate shoot detection in natural plantation scenes. The model combines a channel-spatial feature enhancement bottleneck, a normalized Wasserstein distance constraint for bounding-box regression, and ghost convolution layers in the detection head. We evaluated the model through detector comparisons, module ablations, and repeated training with five matched random seeds on a natural-scene dataset containing four Wuyi rock-tea cultivars. Across the five matched seeds, the mean mAP@0.5 increased from 67.37 ± 0.91% to 67.97 ± 0.90% on the validation set and from 61.29 ± 0.40% to 61.88 ± 0.66% on the internal test set. Neither paired difference was statistically significant: The 95% confidence intervals included zero, and the exact two-sided paired-permutation p values were 0.375 and 0.250, respectively. The mean mAP@0.5:0.95 did not improve. YOLO11s-CSNG retained a model size and model-only edge-inference time comparable to YOLO11s, providing a compact design for candidate shoot-region detection under the sampled field conditions.
Why it matches plant phenotyping methods茶芽という植物器官の画像検出手法を開発し、比較・アブレーション・反復検証を行っているため、植物フェノタイピング手法が中心である。
abstractHere, we developed YOLO11s-CSNG for candidate shoot detection in natural plantation scenes.
Abstract In peanut ( Arachis hypogaea L.), plant stand establishment, seedling vigor, and canopy growth are key determinants of crop performance; yet traditional ground‐based assessment methods can be destructive, labor‐intensive, and limited in throughput. This study evaluated the potential of vegetation metrics derived from unmanned aerial vehicle (UAV)‐based red‐green‐blue (RGB) and multispectral (MS) imagery for high‐throughput, nondestructive assessment of plant stand establishment, seedling vigor, and light interception in peanut. Six runner‐type peanut cultivars were evaluated in 2024 and seven in 2025, with each cultivar represented by two seed size classes (small and large), to generate variation in these traits. Within‐row vegetation discontinuity‐based plant stand ratings for estimating plant stand count ( R 2 = 0.81–0.90), together with canopy coverage for assessing seedling biomass ( R 2 = 0.77–0.82) and light interception ( R 2 = 0.96–0.98), were the best‐performing vegetation metrics. These vegetation metrics provided similar or greater cultivar separation compared with ground‐based measurements. In contrast, several vegetation indices exhibited strong correlations with ground‐based measurements but provided inconsistent cultivar rankings and statistical groupings. MS imagery outperformed RGB imagery for plant stand and seedling biomass assessment. Overall, these results demonstrate that UAV‐derived canopy metrics provide reliable, high‐throughput tools for early‐ to mid‐season crop assessment and offer scalable alternatives to traditional ground‐based approaches for agronomic, crop physiological, and plant breeding research.
Why it matches plant phenotyping methodsUAV画像から植物体の出芽・苗勢・バイオマス・光 interception を推定する植 phenotyping 手法を開発・評価しており、取得指標の性能検証が研究の中心です。
abstractThis study evaluated the potential of vegetation metrics derived from unmanned aerial vehicle (UAV)‐based red‐green‐blue (RGB) and multispectral (MS) imagery for high‐throughput, nondestructive assessment of plant stand establishment, seedling vigor, and light interception in peanut.
Automated plant identification based on leaf morphology has gained significant attention in recent years due to its wide range of applications in precision agriculture, biodiversity conservation, environmental monitoring, and botanical informatics. Advances in digital image processing and machine learning have enabled the development of intelligent systems capable of identifying plant species from leaf characteristics with minimal human intervention. Despite these advancements, achieving reliable and accurate classification remains challenging because leaf images are often affected by variations in illumination, complex backgrounds, image noise, differences in orientation and scale, as well as natural leaf deformation. These factors can obscure important morphological features, reduce the effectiveness of feature extraction, and ultimately decrease the accuracy and robustness of automated plant classification systems. Consequently, there is a growing need for intelligent frameworks that can effectively handle these challenges while preserving critical leaf morphology and venation information for reliable plant identification. This study proposes an Intelligent Morphology-Driven Framework that integrates advanced digital image processing and machine learning for robust leaf venation analysis and plant classification. The proposed framework integrates multiple digital image processing and machine learning techniques to enable accurate and automated leaf venation analysis and plant classification. Initially, leaf images undergo preprocessing using grayscale conversion, histogram equalization, Contrast Limited Adaptive Histogram Equalization (CLAHE), Gaussian filtering, Laplacian sharpening, Gabor filtering, and homomorphic filtering to improve image quality and enhance venation and structural details. The enhanced images are then processed through threshold-based segmentation followed by morphological operations, including erosion, dilation, opening, closing, convex hull generation, and skeletonization, to accurately isolate leaf regions while preserving their geometric structure.To characterize leaf morphology, the framework extracts a comprehensive set of features, including geometric descriptors such as area, perimeter, circularity, aspect ratio, solidity, eccentricity, and vein density, together with Hu invariant moments that provide rotation-, translation-, and scale-invariant shape representation. In addition, the framework investigates the influence of image compression by comparing lossless PNG and lossy JPEG formats to evaluate their impact on preserving morphological features and venation details. The extracted feature vectors are subsequently classified using a Random Forest classifier to categorize leaf venation patterns into parallel, reticulate-pinnate, and reticulate-palmate classes.Experimental evaluation demonstrates that the proposed framework achieves an overall classification accuracy of 93.2%, while effectively preserving important morphological characteristics and maintaining computational efficiency. The combination of adaptive image enhancement, morphology-preserving segmentation, comprehensive feature extraction, and robust machine learning classification makes the proposed approach reliable, interpretable, and scalable. Consequently, the framework has significant potential for applications in digital herbarium systems, automated plant identification, biodiversity monitoring, botanical informatics, and precision agriculture.
Why it matches plant phenotyping methods葉画像から形態・葉脈形質を抽出し分類する画像処理・機械学習フレームワーク自体が研究の中心であり、植物表現型の取得・解析手法として適格。
abstractThis study proposes an Intelligent Morphology-Driven Framework that integrates advanced digital image processing and machine learning for robust leaf venation analysis and plant classification.
Conventional methods of plant distance and volume measurements are limited by low efficiency, limited spatial coverage, and high measurement error. LiDAR and RGB-D imaging offer cost-effective, precise, and non-destructive techniques for plant distance and volume measurements. This study aimed to measure cabbage height, volume, and distance using LiDAR and RGB-D imaging. The sensors were mounted on a 1.6 kW electric field scouting platform (EFSP) for data collection. Point cloud (PCD) data were collected using LiDAR, whereas data processing, visualization, and measurements were done using commercial software and open-source programming scripts. A total of 20 cabbage plants were analyzed. LiDAR data processing included data frame screening, outlier removal, denoising, voxelization, and generation of 3D PCD density maps. Depth image processing included importing raw data and metadata shaping using intrinsic camera parameters, visualization, extraction of depth points, and pixel-level measurements of distances and volume. RGB image processing involved image conversion, segmentation, normalization, binary masking, mask cleaning, region extraction of cabbages, separation of ROI and preparation of contours, Delaunay triangulation and convex hull preparation, ROI overlay, bounding box preparation, sharing boundary between two boxes, conversion to pixel distances, and for visualization, plant height, volume measurements, and center to center distance measurement for measuring the plant distance. LiDAR demonstrated higher measurement accuracy for cabbage plant height, circumferential volume (geometric canopy volume), and plant distance, followed by RGB-D imaging, while RGB imagery showed comparatively lower performance under the study field conditions. Overall, LiDAR and RGB-D imaging provided reliable and non-destructive approaches for cabbage geometric characterization under field conditions, although accurately capturing complex plant geometry remains challenging. Positive and negative values of bias represent the over- and under-estimated results, respectively. Future studies should include larger and more diverse plant datasets exhibiting diversified size, shape, and geometric structure to further improve the robustness and general applicability of the proposed sensing approaches.
Why it matches plant phenotyping methodsLiDAR、RGB、RGB-Dを用いてキャベツの高さ・体積・株間距離を取得し、処理手順と測定精度を比較評価する手法研究であり、植物表現型取得が中心である。
abstractThis study aimed to measure cabbage height, volume, and distance using LiDAR and RGB-D imaging.
Corn is a staple crop of global significance; however, foliar diseases may lead to 30–60% yield loss if not detected at an early stage. Conventional visual inspection is time-consuming, subjective, and difficult to scale for smallholder farmers worldwide. To overcome these issues, we present Corn Transfer Learning Network (CTL-Net), an end-to-end hybrid deep learning model for corn leaf disease identification. CTL-Net, which combines Inception-ResNet-v2 as the backbone and MobileNetV3 as a feature extractor in parallel convolutional streams, simultaneously learns diverse scales of texture information from low-level textures, mid-level structural patterns, and high-level disease semantics of RGB leaf images. Adaptive feature fusion is formulated through learnable weighting coefficients and bi-directional spatial–channel attention mechanisms, which further enhance feature discriminability and robustness. The proposed approach is tested on an extensive dataset of 12,456 images from 10 corn diseases, including Northern Leaf Blight, Common Rust, Gray Leaf Spot, and Cercospora Leaf Spot, acquired under controlled and real-field conditions. CTL-Net attains the highest classification accuracy of 99.42%, outperforming DenseNet121 (97.92%), EfficientNetB3 (97.35%), and general stacking models (97.89%). Robustness experiments demonstrate the effectiveness of the proposed method against illumination variations, additive noise, and partial occlusions. CTL-Net enables real-time inference with a latency of 42 ms on an NVIDIA RTX 3090 GPU. Gradient-weighted Class Activation Mapping++ (Grad-CAM++)-based interpretability analysis results in a mean Intersection over Union (IoU) of 87.6% with expert-annotated disease regions. Five-fold cross-validation, ablation studies, and statistical significance testing (p
Why it matches plant phenotyping methodsトウモロコシ葉画像から病徴・病害状態を推定する深層学習手法を開発し、複数モデルとの比較、頑健性評価、交差検証、アブレーションを行っており、植物フェノタイピング手法が中心である。
abstractwe present Corn Transfer Learning Network (CTL-Net), an end-to-end hybrid deep learning model for corn leaf disease identification.
Reproduction assets foundThe paper states its final curated corn leaf disease dataset (12,456 images, 10 classes) is publicly available via the authors' GitHub repository vishruthkp/maizedataset (reference [30] and Data Availability statement). The Kaggle PlantVillage and Corn or Maize Leaf Disease datasets are cited source inputs, not paper-Dataset · public"Prediction of Crop Yield using Machine Learning," International Available: https://github.com/vishruthkp/maizedataset.Open asset ↗vishruthkp/maizedatasetpdf-page:7 lines:1-53Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
Germination percentage is an endpoint measure and therefore does not describe when an individual seed begins visible growth or how rapidly its radicle and plumule expand. We developed a time-resolved phenotyping workflow to quantify rice seed germination continuously in shallow-water culture. A single industrial camera moved along a 1 m rail and imaged three culture boxes at 1 h intervals for up to 80 h. The archive comprised 1,062 full-frame images and 6,372 seed-level repeated observations under the six-seed field-of-view configuration. A physical grid maintained seed identity through time and enabled individual regions of interest to be extracted. Whole-seed foregrounds were obtained with a pretrained U 2 -Net, and a masked RGB intensity rule separated newly emerging tissue from the darker hull. For each tracked seed, projected emerging-tissue area and interval growth rate were calculated. Three representative normally germinating seeds first showed measurable tissue at 48 h, yet subsequently followed distinct trajectories: final projected areas ranged from 2,605 to 4,700 pixels and peak interval growth rates ranged from 106.88 to 287.92 pixels h −1 . B-1 accumulated 63.71% of its final visible area during 72–80 h, whereas B-3 accumulated 73.51% during 60–72 h. Thus, seeds with the same observed emergence interval can differ substantially in the timing and magnitude of post-emergence expansion. The workflow converts repeated images into biologically interpretable temporal phenotypes and provides a basis for nondestructive studies of rice seed vigor and germination heterogeneity.
Why it matches plant phenotyping methods連続画像から個々のイネ種子の発芽・組織面積・成長速度を抽出する時間分解フェノタイピングワークフローを開発しており、表現型取得と解析手法が研究の中心である。
abstractWe developed a time-resolved phenotyping workflow to quantify rice seed germination continuously in shallow-water culture.
Field / plotMultimodalNeRF / 3D Gaussian SplattingLiDAR / point cloudRGB / grayscaleStereoFruitWhole plant / canopy / plot / fieldMorphology / geometry measurementObject detection
Binocular stereo vision is a low-cost and scalable 3D perception technology that shows strong potential in agricultural phenotyping and smart agriculture. By estimating depth from multi-view RGB images, it enables non-contact, high-precision sensing of crop structure, canopy morphology, growth dynamics, and livestock traits, providing essential support for digital and intelligent agricultural production. With recent advances in deep learning-based stereo matching, multimodal sensor fusion, and 3D reconstruction, its robustness and accuracy in complex field environments have been significantly improved. This paper systematically reviews recent progress in agricultural applications of binocular stereo vision, covering system architectures, traditional and deep learning-based stereo matching methods, point cloud reconstruction techniques, and emerging supervision strategies such as 3D Gaussian splatting. It further summarizes key applications, including high-throughput phenotyping, fruit localization and robotic harvesting, weed detection and precision spraying, autonomous navigation, and livestock body condition assessment, highlighting its role in multi-task agricultural perception systems. Finally, the paper discusses major challenges, including low-texture matching difficulty, occlusions in complex environments, cross-domain generalization, real-time lightweight deployment, and limited dataset availability. Future directions are outlined in foundation model-based visual perception, self- and weakly supervised learning, multimodal fusion, and edge-efficient model design, aiming to support large-scale deployment in smart agriculture.
Why it matches plant phenotyping methods農業における双眼ステレオビジョンのシステム、ステレオマッチング、3D再構成を体系的にレビューし、作物構造・群落形態・生育動態の非接触計測とハイスループット表現型解析を主要対象としているため。
abstractThis paper systematically reviews recent progress in agricultural applications of binocular stereo vision, covering system architectures, traditional and deep learning-based stereo matching methods, point cloud reconstruction techniques
Herein, we rationally designed and synthesized two types of D-π-A red-emitting fluorescent probes with intrinsic intramolecular charge transfer (ICT) and twisted intramolecular charge transfer (TICT) characteristics, for traditional methods for lipid droplet detection in food crops suffer from reliance on bulky instruments, tedious sample pretreatment and the impossibility of on-site quantification. To modulate the molecular skeleton, methoxy groups were introduced, and highly specific recognition of lipid droplet microenvironments was realized through the probes' synergistic spectral response to polarity and viscosity. Typical turn-on fluorescence responses and favorable lipophilicity were displayed by both probes, thus enabling accurate localization of the hydrophobic domains of lipid droplets. Possessing ratiometric fluorescence performance, the modified probe OCH 3 -Phe-FCN efficiently alleviated signal interference from complex matrices, and toward oleic acid was a lower limit of detection presented. By combining smartphone-based RGB colorimetry and image grayscale analysis was a portable detection platform constructed, with fluorescence signals employed to accomplish visual analysis of lipid droplets in food crops. Satisfactory biocompatibility and targeting capability were validated for the as-prepared probes via cell imaging and cytotoxicity evaluations, and far superior imaging quality and signal-to-noise ratio were demonstrated by the modified probe. With the aid of density functional theory (DFT) calculations, systematically elucidated was the spectral mechanism of excited-state transitions synergistically modulated by the two microenvironmental factors. Not only does this work deepen the insight into the response mechanism of dual-responsive fluorescent probes, but also a promising technical strategy is offered for in situ spectral detection of lipid droplets in food crops.
Why it matches plant phenotyping methods食品作物の脂質滴を蛍光プローブとスマートフォン画像解析で可視化・定量する検出法および携帯型プラットフォームの開発が中心であり、植物の細胞状態を測定する技術的貢献が明確です。
abstractBy combining smartphone-based RGB colorimetry and image grayscale analysis was a portable detection platform constructed, with fluorescence signals employed to accomplish visual analysis of lipid droplets in food crops.
This study investigates the automatic classification of crop phenological stages from low-altitude UAV RGB imagery. The dataset included 11,489 images of five crops: sunflower, rapeseed, soybean, wheat, and barley. The images were annotated using the Biologische Bundesanstalt, Bundessortenamt und Chemische Industrie (BBCH) scale, with labels corresponding to either single stages or stage ranges to reflect heterogeneous field conditions and transitional crop states. A pretrained ResNet18 model was adapted to the task using transfer learning. Training was conducted in two stages: first, the classification head was optimized while the backbone remained frozen; second, the entire network was fine-tuned. The model achieved strong internal test accuracy across all crops, with 100% test accuracy for rapeseed and barley, more than 99% for the remaining crops, and a mean accuracy of 99.73% under the studied survey conditions. The results also compare favorably with previously reported studies on UAV-based phenological classification. Overall, the findings support the potential of low-altitude UAV imagery and deep learning for localized phenological assessment of selected field zones in precision agriculture, while broader deployment requires validation across independent fields, seasons, regions, and survey conditions.
Why it matches plant phenotyping methodsUAV画像と深層学習により作物の生育・フェノロジー段階を自動推定する方法が研究の中心であり、植物状態の抽出性能も評価している。
abstractThis study investigates the automatic classification of crop phenological stages from low-altitude UAV RGB imagery.
Unmanned aerial vehicle (UAV) imagery can support plot-scale crop phenotyping, but spectral, RGB and structural predictors may contribute differently to different traits. We compared six predefined feature groups for predicting soybean SPAD and plant height (PH) in a 1.3 ha field experiment in Sanya, China. The field contained 6197 soybean planting plots, of which 234 had paired SPAD and PH measurements. Multispectral bands, vegetation indices (VIs), RGB descriptors and digital surface model (DSM) metrics were extracted from DJI Mavic 3 Multispectral imagery. Six regression algorithms were evaluated using random fivefold cross-validation, spatial block cross-validation and nested spatial cross-validation. Under random cross-validation, ExtraTrees with multispectral bands, VIs and RGB descriptors produced the numerically highest SPAD performance (R2 = 0.589; RMSE = 6.66), while BayesianRidge with multispectral bands, VIs and DSM metrics produced the highest PH performance (R2 = 0.760; RMSE = 7.14 cm). Nested spatial cross-validation yielded R2 = 0.473 and RMSE = 7.56 for SPAD and R2 = 0.690 and RMSE = 8.13 cm for PH. G4 was selected in four of the five outer folds for SPAD, although the selected algorithm varied, and G5 was selected in all five outer folds for PH. VIs improved prediction of both traits relative to the original bands. Adding RGB descriptors produced only a small and model-dependent improvement for SPAD, whereas adding DSM metrics produced a larger and more consistent improvement for PH. The complete feature set did not outperform G4 for SPAD or G5 for PH. The retained models were applied to all 6197 plots to map SPAD, PH and their field relative combinations. Because all of the validations used one field and one UAV acquisition date, the results describe performance within this experiment and do not establish transferability to other sites, years or growth stages.
Why it matches plant phenotyping methodsUAVマルチスペクトル・RGB・構造特徴からSPADと草丈を推定する特徴抽出および回帰手法を、複数の空間交差検証で比較・評価しており、植物表現型取得が研究の中心である。
titleTrait-Specific Contributions of UAV Multispectral, RGB and Structural Features to Soybean SPAD and Plant Height Phenotyping
Field / plotRGB / grayscaleLeafSegmentationGrowth / development / phenology
Bud flushing timing is a crucial attribute for forest geneticists and physiologists studying adaptive variation in tree species. In clonal populations such as Norway spruce seed orchards, accurate assessment of bud flushing is essential, as this attribute strongly influences survival and fitness. However, traditional bud-flushing assessment relies on visual scoring scales that are inherently subjective and susceptible to observer bias, leading to inconsistencies when multiple technicians collect data. To address these limitations, we introduce a novel quantitative phenotype derived from segmented RGB images for quantifying bud flushing that objectively measures the ratio of light-green current-year needles to the total visible needle area of each grafted ramet. By extracting and summarizing digital image data for each ramet, this approach provides a standardized and reproducible assessment of bud flushing phenology. We evaluated clonally replicated ramets in two sites across two seasons and compared this image-derived flushing ratio with conventional visual scoring and color indices. Using mixed models, we quantified genetic control and cross-orchard stability of clonal performance. The image-derived flushing ratio tracked visual-score progression but showed higher broad-sense heritability (0.54–0.70) than visual scoring (0.45–0.65) and canopy color indices (0.20–0.51). Although site means differed strongly, clonal ranking remained stable between orchards for the image-derived flushing ratio and for visual scoring, whereas color indices showed lower stability and greater uncertainty. Overall, these results support the image-derived flushing ratio as a structure-based phenotype that improves the precision and cross-site transferability of genetic evaluation of bud flushing in Norway spruce seed orchards.
Why it matches plant phenotyping methodsRGB画像のセグメンテーションから bud flushing の定量形質を開発し、従来の視覚評価や色指数と比較検証しており、植物表現型取得法が研究の中心である。
abstractwe introduce a novel quantitative phenotype derived from segmented RGB images for quantifying bud flushing
Reliable traits are needed for identification of tea ( Camellia sinensis ) cultivars, yet the stability of leaf morphology and color across leaf positions remains unclear. This study evaluated inter-cultivar variation and positional stability in leaf morphological, RGB color, and SPAD traits in six predominant cultivars. One-year-old shoots were sampled in a completely randomized design, and five fully expanded leaves below the apical bud were analyzed. SPAD values were measured with a chlorophyll meter, and scanned images were used to extract contour and RGB traits. Data were analyzed using ANOVA, correlation analysis, PCA, and discriminant analysis. Leaf morphology differed among cultivars and leaf positions, with significant cultivar-by-position interactions; however, the width-to-length ratio differed among cultivars but remained stable across positions in these cultivars. SPAD values increased with leaf position and were strongly associated with RGB components, being negatively correlated with R and G and positively correlated with B. Morphological traits explained 52.988% of total variance in PCA and yielded 64.6% overall classification accuracy, with LaoHan showing the highest accuracy (83.3%). Misclassification was concentrated among genetically similar cultivars. These findings suggest that stable leaf shape proportions and SPAD-RGB relationships provide useful descriptors, whereas genetic relatedness limits morphology-based cultivar identification under the present conditions.
Why it matches plant phenotyping methods茶品種識別のため、葉の形態・RGB・SPAD特性の取得と安定性、分類性能を中心に評価しており、画像由来形質抽出を含む実質的な表現型解析である。
abstractscanned images were used to extract contour and RGB traits
Reproduction assets found保存済みの本文根拠を更新済みルールで再検証し、公開資産1件を確認しました。Dataset · publicThe following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/biology15151283/s1 , Table S1: Original data of leaf morphological traits, RGB values, and SPAD values from six tea cultivars in this study.Open asset ↗lines:368-409Code / dataset availability confirmedarXiv · OpenAlex · checked 15 Sept 2026
Accurate 3D crop monitoring underpins data-driven precision agriculture by enabling field-scale analysis of plant structure, growth dynamics, and management response. Modern 3D reconstruction methods perform strongly on generic benchmarks, but rendered appearance may not translate into metrically and agronomically useful geometry in crop fields. We introduce UAV3DCrop, a public benchmark of repeated multi-angle unmanned aerial vehicle (UAV) crop surveys. It contains 88,830 RGB images at $5280 \times 3956$ pixels, with a ground sampling distance of 3.6-5.8 mm, from 91 scenes spanning corn, soybean, wheat, and oat. Track A evaluates seven scene-optimized methods -- Neural Radiance Field (NeRF) and 3D Gaussian Splatting (3DGS) variants -- on held-out views, photogrammetry-referenced depth, and canopy-height recovery. Track B tests four pretrained feed-forward models on zero-shot camera-pose and geometry estimation. The scene-optimized methods rank differently across the three targets: Splatfacto-big leads appearance, whereas Scaffold-GS leads depth and is statistically tied with Splatfacto for canopy height. Among feed-forward models, MapAnything leads on seven of the eight metrics, while the remaining models vary more across crops and fail severely on absolute scale in a way that alignment conceals. Repeated acquisitions reveal further sensitivities that differ by output type and by model, associated with position within the acquisition sequence and with tie-point multiplicity. Current 3D reconstruction methods are therefore not yet interchangeable for agronomic use: no single method wins on appearance, geometry, and canopy height at once, and only one of four feed-forward models recovers usable metric scale. The dataset is publicly available at https://link-dev.github.io/UAV3DCrop/
Why it matches plant phenotyping methods植物キャノピー高さという明示的な形質を対象に、UAV 3D再構成手法をベンチマークし、公開データセットとして提供しているため、フェノタイピング手法が中心である。
abstractWe introduce UAV3DCrop, a public benchmark of repeated multi-angle unmanned aerial vehicle (UAV) crop surveys.
Reproduction assets foundThe paper introduces UAV3DCrop, a public benchmark of repeated multi-angle UAV crop surveys (88,830 RGB images, 91 scenes, four crops) with refined poses, photogrammetric depth references, and linked canopy-height and effective-LAI field measurements. The dataset is explicitly stated to be publicly available under CC BDataset · publiche acquisition sequence and with tie-point multiplicity. Current 3D reconstruction methods are therefore not yet interchangeable for agronomic use: no single method wins on appearance, geometry, and canopy height at once, and only one of four feed-forward models recovers usable metric scale. The dataset is publicly available at https://link-dev.github.io/UAV3DCrop/ .
Keywords:
UAV imagery; agricultural datasets; crop-field reconstruction; neural radiance fields;
Gaussian splatting; feed-forward geometry.
1 IntroductionOpen asset ↗UAV3DCroplines:1-90Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · Europe PMC · checked 5 Sept 2026
Accurate identification and effective prediction of the maize tasseling stage are of great significance for guiding precision field management and ensuring stable crop yields. Conventional manual observation methods suffer from high labor intensity, poor timeliness, and strong subjectivity. In this study, based on an unmanned aerial vehicle (UAV) remote sensing platform, LiDAR point cloud data and RGB imagery were simultaneously acquired to construct digital surface models (DSMs) and digital terrain models (DTMs). Multi-dimensional statistical features were extracted to establish a high-precision plant height estimation method applicable to the entire growth cycle of maize. On this basis, the Logistic growth curve function was introduced to fit the dynamic changes in plant height, enabling the identification and early prediction of the maize tasseling stage based on the plant height growth curve. The research results indicate the following: (1) For maize plant height estimation, the LiDAR sensor outperforms RGB. The optimal accuracy is achieved by combining the 99th percentile of DSM with the minimum DTM, yielding a root mean square error (RMSE) of 0.17 m. (2) Based on the high-accuracy plant height time series, the point of inflection (POI) achieves the highest accuracy in tasseling stage identification, with an RMSE of 2.586 d under the reconstructed time series. (3) Prediction accuracy of the tasseling stage improves with increasing plant height threshold, and optimal performance is observed when the threshold is ≥1.6 m with a growth rate between 0.11 and 0.13. This study establishes a technical framework of "time-series perception-dynamic simulation-feature identification-early prediction", providing a scientific basis for automated monitoring and precision management of the maize tasseling stage. It holds significant theoretical and practical value for the advancement of smart agriculture and crop phenotyping research.
Why it matches plant phenotyping methodsUAV LiDAR/RGBによる草丈推定と時系列成長曲線から、トウモロコシの抽だい期を識別・予測する技術フレームワークが研究の中心であり、植物形質取得と検証を伴うため採用。
abstractMulti-dimensional statistical features were extracted to establish a high-precision plant height estimation method applicable to the entire growth cycle of maize.
Plant height is a key 3D phenotypic trait for assessing crop growth, biomass accumulation, and lodging resistance. To overcome the practical limitations of conventional plant height measurement methods, this study proposes a novel vision foundation model-based framework named Depth for Plant Height (Depth4PH) for plant height estimation in agricultural scenes using low-cost monocular RGB imaging. As part of our contributions, a synthetic–real coupled multimodal dataset was constructed by integrating Blender virtual agricultural scenes (Blender VAS) with real field images. Building upon the existing Depth Anything V2 foundation model, we developed a novel module called Transfer-based Agricultural Metric Depth Anything V2 (TAM-Depth V2) for absolute metric depth estimation through parameter-efficient fine-tuning, depth decoder reconstruction, and joint loss optimization. Furthermore, we designed a novel multi-source prompt-based segmentation framework, MSP-SAM2, to generate positive and negative prompts for zero-shot crop instance segmentation. Finally, a new inverse physical plant height estimation algorithm, RANSAC-Per, was introduced to estimate plant height by combining truncated percentile statistics with local RANSAC micro-plane fitting, thereby reducing the effects of depth noise and field microtopographic variation. The result showed that TAM-Depth V2 achieved stable absolute depth estimation, with an RMSE of 0.1162 m and an AbsRel of 4.25%. Compared to the original box-prompted SAM 2, MSP-SAM2 achieved a 4.4% improvement in mIoU, reaching 91.6% and a recall of 93.2%. On a 350-plant multi-crop test set, Depth4PH achieved R 2 = 0.948, RMSE = 12.23 cm, and MAE = 8.82 cm, and MAPE =10.15%, with crop-specific RMSEs ranging from 3.99 cm (cucumber) to 20.73 cm (maize), significantly outperforming the traditional Global-MinMax baseline (which had an RMSE of 22.62 cm). These results indicate that Depth4PH provides a promising foundational pathway for high-throughput crop phenotyping. With future optimization for edge deployment, it holds significant potential to support high-throughput monitoring in precision agriculture.
Why it matches plant phenotyping methods植物高の画像取得・深度推定・セグメンテーション・高さ抽出アルゴリズムを一体化した植物表現型計測フレームワークの開発と検証が中心であり、データセット構築と性能評価も含む。
abstractthis study proposes a novel vision foundation model-based framework named Depth for Plant Height (Depth4PH) for plant height estimation in agricultural scenes using low-cost monocular RGB imaging.
Spike count and spatial arrangement are among the strongest determinants of grain yield in wheat, yet reliable spike-level phenotyping under field conditions remains challenging. Field-acquired images are often affected by occlusion, heterogeneous illumination, and dense canopy overlap. In addition to these visual complexities, the effort required to generate large-scale instance-level annotations makes it difficult to build the extensive labeled datasets typically needed for robust segmentation models. This paper addresses both problems. The core methodological contribution is a semi-automated iterative annotation workflow: a YOLOv11x-seg model trained on a small manually annotated set is applied to unlabeled field images, and its predictions, after post-processing to remove duplicated, fragmented, and merged detections, are incorporated back into training. To prevent augmentation from distorting the training distribution, we introduce a distribution-aware augmentation strategy guided by Fréchet Inception Distance (FID), which retains only those augmented samples that remain within an acceptable distance from the original data distribution. Together, these components allowed us to build an effective training set from 3,899 high-resolution RGB images (4000 × 3000 pixels) of durum wheat collected at the CREA Research Centre for Cereal and Industrial Crops, with substantially reduced manual annotation effort. On the independent test set(242 images), the final YOLOv11x-seg model (M5 model) achieved a mask-level precision of 86.73%, recall of 83.02%, F1-score of 84.83%, mAP@50 of 89.42%, and mAP@50:95 of 60.51%. Spike masks were used to derive image-based traits including spike count, spike density, canopy coverage, spike area, spatial distribution, and vegetation indices. Their relationships with measured grain yield were explored through statistical analysis and machine-learning-based yield estimation.Both statistical and machine-learning analyses demonstrated that image-derived spike traits provided meaningful information for grain yield estimation, with canopy coverage showing the strongest positive association with yield. Using repeated nested cross-validation with out-of-fold (OOF) predictions, XGBoost achieved the highest yield estimation performance ( R OOF 2 = 0.312 , RMSE = 106.61 g/plot), supporting the potential of near-image phenotyping for late-stage yield estimation in wheat. These results show that semi-automated iterative annotation can enable practical wheat spike segmentation and image-based phenotyping under realistic open-field conditions. While grain yield estimation should be interpreted within the context of the experimental setting, the proposed framework highlights the value of image-derived spike traits for late-stage phenotyping and yield assessment in breeding experiments rather than for early-season yield forecasting.
Why it matches plant phenotyping methods半自動アノテーション、YOLOv11x-segによる小麦穂の画像セグメンテーション、分布認識型データ拡張、形質抽出と検証が中心的な方法論的貢献である。
abstractThe core methodological contribution is a semi-automated iterative annotation workflow
Finger millet is an important cereal crop widely cultivated worldwide for food and fodder. Breeding programs aim to select genotypes with desirable architectural traits to develop new varieties with higher yields. In this effort, accurate high-throughput plant phenotyping is essential for accelerating crop improvement. To overcome the time-consuming and labor-intensive process of manual measurements, this study presents a comprehensive 3D imaging pipeline that leverages neural radiance fields (NeRF), 3D gaussian splatting (3DGS), and its advanced extensions (e.g., Feature 3DGS and Gaussian Grouping) to reconstruct, segment, and analyze finger millet yield component traits using multi-view 2D images. First, multiple-view RGB images of a single finger millet plant were captured, and COLMAP was then utilized to estimate the camera poses of the images and reconstruct the sparse point cloud, followed by advanced 3D reconstruction through 3DGS and NeRF. Second, feature 3DGS and gaussian grouping models were used to generate the 3D gaussian representation of finger millet panicles. This single-process framework enabled the generation of high-fidelity 3D point clouds and semantic feature fields without the need for expensive depth sensors or manual annotations. Our results demonstrated the effectiveness of these models in capturing morphological variations across different panicle phenotypes, including compact versus open panicle architectures. In addition, the 3D point clouds of the panicles were utilized to extract structural traits for yield prediction, achieving biologically meaningful correlations with grain productivity. This work highlights the potential of 3DGS-based phenotyping pipelines as a low-cost, near real-time, photorealistic solution for trait quantification, segmentation, and yield estimation in real-world agricultural settings.
Why it matches plant phenotyping methods3D画像再構成・セグメンテーション・形質抽出を統合した植物フェノタイピング手法の開発が中心であり、収量関連形質の定量と予測まで技術的に評価している。
abstractthis study presents a comprehensive 3D imaging pipeline that leverages neural radiance fields (NeRF), 3D gaussian splatting (3DGS), and its advanced extensions (e.g., Feature 3DGS and Gaussian Grouping) to reconstruct, segment, and analyze finger millet yield component traits using multi-view 2D images.
• Development of yellow color index (YCI) for yield estimation at flowering stage • Development of web-based interface (YCPM-UAV) for canola yield prediction using UAVs • Global application capability for UAVs datasets to predict canola yield using YCPM-UAV • Multi-sensor and multi-spectrum data fusion to find most suited indices for canola • Multiple and stepwise regression analysis for selection of most influencing VIs Canola ( Brassica napus L.) is a globally significant oilseed crop, yet accurate yield estimation remains challenging due to the complex and unique nature of the crop, especially at the flowering stage. Traditional field-based yield estimation methods are labor-intensive, time-consuming, and destructive, necessitating innovative approaches for early and non-destructive yield prediction. The main objective of the study is to develop a novel web-based platform, YCPM-UAV (Yellow Color Prediction Model using Unmanned Aerial Vehicles), for early and accurate canola yield estimation using high-resolution multi-sensor datasets acquired through low-altitude UAVs (LA-UAVs). To achieve this objective, a comprehensive two-year field study (2022-2024) was conducted across ten farmers’ fields in different geographical locations. Multisensor data (RGB, multispectral, and thermal) were acquired using UAVs at seven growth stages. Several vegetation indices (VIs), yellow color-based indices, and a thermal index were calculated. Linear, multiple, and stepwise regression analyses were performed to evaluate relationships of remote sensing indices with ground-truth yield data collected from 1200 sampling points. Multiple and stepwise regression analyses indicated that the newly developed Yellow Color Index (YCI) exhibited the strongest correlation with actual canola yield at the flowering stage across both years (Year 1: R 2 = 0.84, RMSE = 39.30 g m⁻²; Year 2: R² = 0.88, RMSE = 31.57 g m⁻²). Based on proposed predictive modeling, the YCPM-UAV web interface was developed, featuring automated data processing and spatial analysis with a testing accuracy of 88%. The YCPM-UAV platform provides farmers, researchers, and policymakers with a timely, user-friendly, and actionable decision-support tool for canola yield estimation at the field scale, contributing to improved crop management and food security. Future studies should incorporate additional canola varieties, irrigated and non-irrigated fields, and deep learning algorithms to further improve model robustness.
Why it matches plant phenotyping methodsUAVマルチセンサー画像からカノーラ収量を推定する指標・回帰モデル・Webプラットフォームを開発し、複数年データで検証しており、植物形質取得・推定が中心である。
titleDevelopment of web-based YCPM-UAV interface for early yield prediction of canola crop using UAV multi-sensor data
Rice panicle number per unit area is a key determinant of yield, but manual counting remains time-consuming and labor-intensive. This study proposes CEG-YOLO, a lightweight deep learning model for in-field rice panicle detection and counting using consumer-grade RGB imaging devices. The model introduces three improvements to YOLOv11s to address specific challenges in field scenarios: C2f-Fast replaces standard convolutions with depthwise convolutions to reduce computational cost for edge deployment; SPPF-ECA integrates an attention mechanism to suppress complex background interference; and GhostConv reduces feature redundancy to improve detection of dense panicles. A dataset of 5,175 images was collected from four rice cultivars planted at three densities using an iPhone 12. The proposed model achieved 93.9% average precision (AP) on the test set, outperforming YOLOv11s which achieved 89.1%, while reducing parameters to 7.8 million and floating-point operations (FLOPs) to 16.5 billion. Robustness evaluation yielded coefficients of determination (R²) values of 0.91 to 0.94 for lighting, 0.89 to 0.94 for planting density, and 0.90 to 0.94 for cultivar. A cross-year field test using an NVIDIA Jetson Orin NX edge device on 120 quadrats in 2025 achieved R² of 0.91, root mean square error (RMSE) of 4.0, and mean absolute error (MAE) of 3.3 at 20 frames per second, confirming practical deployability. This study demonstrates that smartphone-based proximal phenotyping with an optimized deep learning model can provide accurate, low-cost rice panicle counting for breeding and production applications.
Why it matches plant phenotyping methodsイネ穂数という植物形質をRGB画像から推定する深層学習モデルを開発し、精度・頑健性・実地展開性能を検証しており、フェノタイピング手法が研究の中心である。
abstractThis study proposes CEG-YOLO, a lightweight deep learning model for in-field rice panicle detection and counting using consumer-grade RGB imaging devices.
This paper presents an IoT and edge-AI enabled autonomous agricultural robot that performs early plant disease diagnosis and precision irrigation on a single mobile platform. Unlike earlier automated farming systems that rely on visible-spectrum (RGB) imagery and simple threshold-based watering, the proposed system fuses RGB and nearinfrared (NIR) imagery to compute the Normalised Difference Vegetation Index (NDVI), enabling detection of physiological plant stress several days before visible lesions appear. Leaf images are classified using a lightweight attention-guided convolutional neural network that combines a MobileNetV3 backbone with a Convolutional Block Attention Module (CBAM), allowing the network to focus on lesion-relevant channels and spatial regions while remaining compact enough for real-time inference on an ESP32-S3 edge controller. Irrigation and pesticide-spray decisions are no longer governed by a rigid binary threshold; instead, a Mamdani-type fuzzy inference engine fuses soil moisture, ambient temperature, and the NDVI-derived stress index to compute a proportional, continuously variable actuation signal, reducing both water wastage and false triggering. The robot streams sensor readings, classification results, and actuation logs to a cloud dashboard over Wi-Fi/MQTT so that farmers can monitor crop health and irrigation status remotely and receive real-time alerts. Experimental evaluation on a prototype platform shows that the proposed attention-guided model improves disease-classification accuracy over a baseline CNN, the NDVI-assisted pipeline detects stress earlier than colour-only analysis, and the fuzzy irrigation controller reduces water consumption relative to the binary threshold scheme while maintaining optimal soil-moisture levels. The results indicate that combining multispectral sensing, attention-based lightweight deep learning, and fuzzy control on a single autonomous platform is a practical and scalable route towards sustainable, resource-efficient precision agriculture.
Why it matches plant phenotyping methodsRGB/NIR画像からNDVIによる植物ストレスを推定し、葉画像から病徴を分類する取得・解析手法をロボット上で開発・評価しており、植物表現型の測定が中心である。
abstractthe proposed system fuses RGB and nearinfrared (NIR) imagery to compute the Normalised Difference Vegetation Index (NDVI), enabling detection of physiological plant stress several days before visible lesions appear.
Quantifying the canopy growth dynamics and light interception capacity under different management practices laid the physiological foundation for potato yield formation. However, the traditional manual measurement methods are labour-intensive, time-consuming, and incapable of capturing time-series dynamics. To address this, we proposed a novel high-throughput strategy that integrates UAV-based RGB imaging with a piecewise physiological model. Furthermore, how Nitrogen(N)-Potassium(K) interaction affects the temporal canopy growth dynamics, light interception, and tuber yield was determined. The results indicated that: (1) Among the 11 secondary indices extracted from the canopy growth dynamic curves, the interaction of N and K had the greatest effect on the maximum canopy duration and the total canopy growth curve integral. The direct path coefficients of N and K inputs on these two parameters were 0.847 and 0.805, and 0.234 and 0.148, respectively. (2) There was a strong linear relationship between the integral area under the curve (S∫) and the total plant dry weight, with R² at 0.90 in 2023-2024. A simplified net photosynthetically active radiation utilisation assessment framework that achieved high accuracy with minimal parameter was built. (3) Prolonging the maximum canopy continuous coverage time is the main way to improve potato yield. The overall effect of N input on yield was significantly higher than that of K fertiliser, with a total effect value of 1.428. Optimising the N-K interaction improves nutrient precision and light interception. The integration of UAV remote sensing and the crop physiological-ecological model enables the tracking of potato canopy dynamics, which is helpful for optimising management practices to improve potato yield.
Why it matches plant phenotyping methodsUAV RGB画像と生理モデルを統合した高スループット手法を開発し、ジャガイモのキャノピー成長動態と光 interception を時系列で推定することが中心である。
abstractTo address this, we proposed a novel high-throughput strategy that integrates UAV-based RGB imaging with a piecewise physiological model.
• DL model trained on MS data achieved the highest accuracy with an R 2 of 75.29% • Linear Regression Coefficient-Based feature selection with SVM and PCA with Linear Regression significantly improved model performance. • NIR and red-edge bands in the MS dataset consistently outperformed the RGB dataset • More represented rice variety (Sona) achieved a strong R 2 of 80.21% on MS data Accurate crop yield prediction is critical for agricultural planning, food security assessment, and farm-level decision-making. In Nepal, however, rice yield estimation is still predominantly based on traditional approaches, where local agricultural extension offices collect field-level observations that are subsequently aggregated at district, provincial, and national scales, often limiting spatial detail and timeliness. This study aims to develop a field-scale rice yield estimation framework by integrating Unmanned Aerial Vehicle (UAV)-derived remote sensing data with machine learning (ML) and deep learning (DL) techniques. High-resolution multispectral (MS) and RGB UAV imagery were used to evaluate the influence of Vegetation Indices (VIs), including HUE and VNDVI from RGB data and RGBVI and Simple Ratio (SR) from MS data, along with plant characteristics and farm management practices (e.g., application of Zyme and Zinc Potash) on rice yield. The predictive performance of Support Vector Machines (SVM), Linear Regression (LR), Decision Trees (DT), Random Forests (RF), and deep neural network models were systematically assessed. Data preprocessing included feature selection based on importance ranking, Yeo–Johnson power transformation, and Principal Component Analysis (PCA) to improve model stability and performance. Among conventional ML models, LR combined with PCA achieved a coefficient of determination (R²) of 69.09% using MS data, while SVM yielded the best performance using RGB data (R² = 68.27%). Overall, deep neural networks outperformed other models, achieving R² values of 75.29% and 64.60% for MS and RGB data, respectively. Model performance varied notably across rice varieties; the Sona variety (n = 127) achieved the highest coefficient of determination (R² = 80.21% for MS and 76.34% for RGB), whereas varieties with fewer samples exhibited lower predictive performance. Results further indicate that ranking features by importance, rather than eliminating them, enhances predictive accuracy, particularly when using LR-derived feature importance, which proved critical for improving the performance of both LR and SVM models.
Why it matches plant phenotyping methodsUAV画像からイネ収量を推定する手法・フレームワークの開発と、複数の機械学習モデルの系統的評価が研究の中心であり、単なる収量のルーチン測定ではない。
abstractThis study aims to develop a field-scale rice yield estimation framework by integrating Unmanned Aerial Vehicle (UAV)-derived remote sensing data with machine learning (ML) and deep learning (DL) techniques.
Field / plotLiDAR / point cloudRGB / grayscaleRootStem / branchWhole plant / canopy / plot / fieldMorphology / geometry measurement2D/3D reconstructionVisualization / data managementArchitecture / morphology / geometry
Diameter at breast height (DBH) is a crucial indicator for obtaining tree phenotypes in orchard management, plantation monitoring, and agroforestry systems. LiDAR technology has high measurement accuracy, but it is costly and difficult to deploy flexibly in outdoor scenarios, while smartphones have emerged as a viable alternative due to their portability and low cost. In this paper, we propose a DBH estimation method based on monocular depth estimation, supported by a mobile application for algorithm deployment and result visualization. To address the limited computing resources on mobile devices, we design HR-DiffusionDepth, a lightweight diffusion-based monocular depth estimation network for smartphones, which generates pixel-wise 3D coordinates from a single image using camera intrinsics, thereby replacing LiDAR for DBH calculation. Experiments on the KITTI and SPREAD datasets show that HR-DiffusionDepth achieves the best depth estimation accuracy among similar lightweight models, reducing Abs Rel by up to 25.3% relative to the state-of-the-art (SoTA) lightweight baseline, with only 6.26 M parameters. The validation results show that the root mean square error (RMSE) of DBH estimation is 3.10 cm and the mean absolute error (MAE) is 2.25 cm, demonstrating the potential of this approach for agricultural scenarios such as orchards and plantations.
Why it matches plant phenotyping methodsスマートフォン画像と軽量深度推定ネットワークにより樹木DBHを推定する手法を開発・検証しており、植物形質取得が研究の中心である。
abstractwe propose a DBH estimation method based on monocular depth estimation, supported by a mobile application for algorithm deployment and result visualization.
With the rapid development of artificial intelligence, UAV remote sensing, and agricultural Internet of Things technologies, crop growth monitoring is evolving from manual inspection and single-source analysis toward intelligent decision-making based on multisource perception. However, existing methods still suffer from limited robustness under environmental variations, insufficient integration between UAV imagery and sparse ground sensor observations, and weak capability for transforming predictions into practical agricultural management recommendations. This study proposes a UAV–ground sensor collaborative lightweight framework for crop growth assessment and agricultural decision support. The proposed framework integrates UAV RGB and multispectral imagery with ground sensor observations through a region-level aerial–ground alignment mechanism and a sensor-guided attention fusion module, enabling environmental conditions to enhance visual feature interpretation. Furthermore, a fact-constrained decision module is developed to generate management recommendations based on crop status, environmental risks, and field information. Experimental results demonstrate that the proposed method achieves superior performance in crop growth classification and yield-trend prediction, reaching Accuracy, Precision, Recall, and F1-score values of 92.47%, 91.86%, 91.39%, and 91.62%, respectively, with an RMSE of 0.381 and an R2 of 0.902. The lightweight framework requires only 6.18M parameters and 0.91G FLOPs, achieving 39.56 ms inference latency and 25.28 FPS on edge devices. The proposed framework also improves decision reliability, achieving an expert agreement rate of 89.34% and a risk identification accuracy of 90.18%. Economic analysis indicates that the proposed framework reduces labor cost, water consumption, and fertilizer input by 49.7%, 26.7%, and 23.0%, respectively, while increasing net benefit by 46.1% compared with conventional field management practices. These results demonstrate that the proposed method provides an accurate, interpretable, and deployable AI-driven solution for intelligent crop management in smallholder and medium-sized farming systems.
Why it matches plant phenotyping methodsUAV画像と地上センサーを融合し、作物生育状態を評価する取得・推定フレームワーク自体を開発しており、植物状態の推定方法が中心的です。
abstractThis study proposes a UAV–ground sensor collaborative lightweight framework for crop growth assessment and agricultural decision support.
Penyakit Basal Stem Rot (BSR) merupakan penyakit penting pada tanaman kelapa sawit yang dapat menurunkan produktivitas. Analisis berbasis citra dapat mendukung deteksi penyakit, tetapi karakteristik distribusi warna perlu dipahami sebelum tahap klasifikasi. Penelitian ini bertujuan menganalisis distribusi warna citra daun kelapa sawit kelas Healthy dan BSR menggunakan Histogram pada ruang warna RGB dan HSV serta mengevaluasi kemampuan diskriminatif fitur warna. Dataset terdiri atas 2.438 citra dari repositori terbuka Roboflow. Setiap citra diproses melalui resize 224×224 piksel, Gaussian Blur, konversi RGB ke HSV, serta ekstraksi 12 parameter statistik berupa mean dan standar deviasi. Perbedaan distribusi antar kelas diuji menggunakan Mann–Whitney U dan besarnya perbedaan dihitung menggunakan Cohen's d. Kemampuan diskriminatif G_Mean, S_Mean, dan V_Mean dievaluasi menggunakan Receiver Operating Characteristic (ROC) dan Area Under the Curve (AUC). Hasil menunjukkan ketiga fitur memiliki perbedaan signifikan (p-value = 0,0000) dengan ukuran efek besar. Nilai AUC G_Mean, S_Mean, dan V_Mean masing-masing sebesar 0,9794, 0,7362, dan 0,9790. Hasil menunjukkan bahwa analisis distribusi warna dapat mengidentifikasi fitur diskriminatif sebagai dasar pemilihan fitur dan perancangan pra-pemrosesan citra BSR.
Why it matches plant phenotyping methods葉の画像からBSR感染状態を色特徴として抽出・評価する画像解析手法が研究の中心であり、単なる病害実験の routine measurement ではない。
abstractAnalisis berbasis citra dapat mendukung deteksi penyakit, tetapi karakteristik distribusi warna perlu dipahami sebelum tahap klasifikasi.
Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · Europe PMC · checked 5 Sept 2026
Background: Rice breeding requires faster development of high-yielding, climate-resilient, resource-efficient, and high-quality cultivars for production systems exposed to environmental variability and increasing input constraints. Genomic selection offers an opportunity to predict breeding value before extensive field evaluation, although its effectiveness depends on the integration of genomic, phenotypic, and environmental information. Methods: This narrative review critically examines recent advances in genomic selection for rice and its integration with high-throughput genotyping, high-throughput phenotyping, machine learning, multi-environment prediction, and speed breeding. Results: Genome-wide marker data can support early ranking of breeding materials for grain yield, grain quality, disease resistance, drought tolerance, salinity tolerance, and nutrient-use efficiency. Prediction performance is influenced by trait architecture, marker density, training-population size, genetic relatedness between training and candidate populations, phenotypic data quality, and genotype-by-environment interaction. Red-green-blue, multispectral, hyperspectral, thermal, and light detection and ranging platforms can generate temporal traits associated with plant architecture, biomass, water status, nutrient status, and stress responses, which may improve prediction under suitable population and validation designs. Speed-breeding systems shorten generation intervals and facilitate rapid advancement, recurrent selection, and recycling of superior parental lines. Conclusions: Integrated breeding pipelines that combine genomic prediction, high-throughput phenotyping, environmental data, and speed breeding can improve selection efficiency and shorten rice improvement cycles. Wider adoption will require affordable technology platforms, standardized data systems, multi-environment validation, breeder capacity development, and collaborative data-sharing frameworks for smart and greener agriculture.
Why it matches plant phenotyping methods高スループット表現型解析をゲノム選抜との統合という方法論的主題の一部として批判的にレビューしており、各種画像・センサープラットフォームと形質抽出を扱うため、表現型手法レビューに該当する。
abstractThis narrative review critically examines recent advances in genomic selection for rice and its integration with high-throughput genotyping, high-throughput phenotyping, machine learning, multi-environment prediction, and speed breeding.
Turfgrass phenotyping relies heavily on visual quality (VQ) ratings and RGB indices like DGCI, but these are limited by observer subjectivity, coarse ordinal scales, or ratio formulations that do not reflect perceptual color differences. Hyperspectral and machine-learning tools overcome some limitations but remain costly and difficult to generalize, motivating the need for scalable and interpretable RGB color metrics. We introduce ΔEg, a perceptually anchored CIELAB ΔE distance from an ideal green that provides a continuous and interpretable measure of canopy color evaluated alongside a panel of RGB-derived metrics. A 3 × 3 nitrogen × irrigation greenhouse experiment using hybrid bermudagrass (TifTuf, Cynodon dactylon × C. transvaalensis) quantified canopy responses with RGB imaging, spectral reflectance, CCM-300 fluorescence, and chlorophyll assays. ΔEg correlated strongly with chlorophyll (r = 0.72), similar to DGCI (r = 0.73), and both exceeded CCM-300 (r = 0.29). HSVi showed the strongest association with VQ (r = 0.84) and was most sensitive to irrigation (ηp2 = 0.63). CIELUV v* explained the greatest model variation (R2m = 0.94) and responded most to fertilizer (ηp2 = 0.84). The yellow fraction was significant across all main and interaction effects and captured canopy decline (r = −0.82 with VQ). An illustrative decision-support scenario using ΔEg indicated that moderate fertilizer combined with mild deficit irrigation optimized turf color and input efficiency. Conclusions apply to controlled conditions, with field-scale validation identified as future work. These results demonstrate that interpretable RGB color metrics, anchored by ΔEg, offer a scalable alternative to VQ scoring and spectral systems.
Why it matches plant phenotyping methodsRGB画像から芝草キャノピー色を定量化するΔEgなどの指標を導入・比較し、クロロフィルや品質評価との技術的関連性を検証しており、植物表現型取得法が中心である。
abstractWe introduce ΔEg, a perceptually anchored CIELAB ΔE distance from an ideal green that provides a continuous and interpretable measure of canopy color evaluated alongside a panel of RGB-derived metrics.
Reproduction assets foundThe paper's Data Availability Statement deposits the phenotype data and the authors' Python image-processing/metric-computation scripts and R statistical analysis scripts in the USDA National Agricultural Library Ag Data Commons, a public repository. The full RGB imagery archive, however, is only available upon requestCode · public2025;23:673–687. doi: 10.1002/lom3.10705.
Associated Data
Data Availability Statement
Data and Python scripts used for image processing and %G, %Gr, %Y, ΔEg, DGCI, HSVi, BA SD , CIELUV v* metric computation, and R scripts used for statistical analysis are be available in the USDA National Agricultural Library Ag Data Commons ( https://agdatacommons.nal.usda.gov/ ), Data for—Proxima Green: RGB Color Metrics for Turfgrass Phenotyping in Controlled Conditions, accessed on 27 July 2026. The full RGB imagery archive will be made available upon reasonable request.Open asset ↗USDA National Agricultural Library Ag Data Commonslines:691-695Plant phenotyping relevance match · UnverifiedOpenAlex · checked 5 Sept 2026
The ready-to-eat lettuce industry is rapidly expanding, increasing the need for reliable, scalable methods to assess seed germination and early growth under realistic soil conditions. This study presents an automated imaging-based approach for quantifying germination dynamics and seedling vigor using a low-cost multi-camera system under greenhouse conditions. Lettuce seeds were grown in soil either inoculated or non-inoculated with the soil-borne pathogen Rhizoctonia solani. Top-view images were acquired using commercial surveillance cameras and processed through a calibrated pipeline including geometric correction, color normalization, vegetation segmentation, clustering, and temporal tracking of emergence events. Seedling vigor was quantified through projected leaf area estimation. The proposed method enables accurate estimation of germination kinetics and growth dynamics under field-like conditions. Automated counts were validated against manual measurements at both intermediate and final time points, achieving high agreement in both cases. At the final assessment, the method reached R² = 0.98 and RMSE = 1.12, while at the midterm evaluation it achieved improved performance with R² = 0.998 and RMSE = 0.5, reflecting the lower complexity of plant structure at earlier growth stages. Results showed that pathogen inoculation significantly reduced both germination rate and seedling vigor, with up to 70% reduction in biomass accumulation. The proposed framework provides a robust, low-cost solution for high-throughput phenotyping of early plant development in soil-based systems, supporting scalable agricultural experimentation.
Why it matches plant phenotyping methods低コスト多カメラ画像システムと画像解析パイプラインを開発・検証し、発芽動態と幼植物活力を定量化しているため、植物フェノタイピング手法が中心です。
abstractThis study presents an automated imaging-based approach for quantifying germination dynamics and seedling vigor using a low-cost multi-camera system under greenhouse conditions.
Abstract Deep learning has achieved remarkable success in rice disease diagnosis; however, existing methods often suffer from limited interpretability and poor robustness against open-world environmental noise. To address these challenges, this study proposes the Knowledge-Guided Multi-Task Rice Network (MTRNet) built upon a ResNet-50 backbone. Unlike conventional "black-box" models, MTRNet employs expert knowledge injection via a phytopathological matrix to explicitly disentangle disease features into Shape, Color, and Location attributes within a multi-head architecture. Furthermore, to mitigate false positives in complex field scenarios, a non-parametric Cascade Inference System (CIS)—comprising a biological grayscale filter and a visual consistency check—is introduced for robust Out-of-Distribution (OOD) detection and anomaly rejection. Experiments on a benchmark dataset of 5,932 field images, which primarily comprises four main rice diseases (Rice Leaf Blast, Brown Spot, Bacterial Leaf Blight, and Tungro), demonstrate that MTRNet achieves a diagnostic accuracy of 99.83%. Crucially, in an open-world robustness evaluation involving 1,000 non-agricultural noise samples, the proposed system achieved an 81.80% OOD rejection rate. By balancing diagnostic accuracy with structural transparency, this framework effectively narrows the gap between laboratory benchmarks and real-world agricultural applications.
Why it matches plant phenotyping methodsイネ病害の画像から病徴を診断する深層学習・OOD検出手法を提案し、実画像データで性能評価しており、植物の病害状態の取得・推定が研究の中心である。
abstractthis study proposes the Knowledge-Guided Multi-Task Rice Network (MTRNet) built upon a ResNet-50 backbone.
Introduction Banana production is increasingly threatened by fungal diseases such as Fusarium wilt and Black Sigatoka, posing severe risks to food security and agricultural economies. Recent image-based approaches using deep learning have shown high predictive capacity for plant disease recognition; however, their limited transparency, calibration uncertainty, and sensitivity to domain shifts can restrict their use in decision-support workflows that require auditability. Methods This study proposes an interpretable and calibrated Artificial Intelligence framework for multiclass banana disease-pattern characterization based on radiomic feature analysis of RGB leaf images. Radiomic features were extracted from HSV-segmented banana leaf regions, resulting in a dataset of 14,763 samples characterized by 103 quantitative descriptors and labeled as Healthy, Sigatoka, or Fusarium wilt race 1. Results Among the evaluated radiomics classifiers, the calibrated Random Forest achieved accuracy = 0.85, balanced accuracy = 0.84, macro-F1 = 0.84, and macro ROC-AUC OvR = 0.95 on the held-out test set. Bootstrap analysis yielded 95% confidence intervals of [0.8406, 0.8691] for accuracy and [0.8306, 0.8606] for balanced accuracy. Three deep learning baselines trained on the same partition achieved higher predictive performance: MobileNetV3 with accuracy = 0.96, macro-F1 = 0.95, and macro ROC-AUC OvR = 0.97; EfficientNet with accuracy = 0.96, macro-F1 = 0.97, and macro ROC-AUC OvR = 0.96; and ResNet-18 with accuracy = 0.96, macro-F1 = 0.96, and macro ROC-AUC OvR = 0.96. Discussion The CNNs produced strong classification performance on the evaluated repositories, and the radiomics approach demonstrated to be a complementary interpretable and explainable calibrated reference model. SHAP, LIME, permutation importance, accumulated local effects, calibration curves, and Brier score decomposition supported feature-level inspection of the final model.
Why it matches plant phenotyping methodsバナナ葉のRGB画像から病害状態を抽出・分類する画像解析および説明可能な機械学習手法が研究の中心であり、植物病害表現型の評価性能も検証している。
abstractThis study proposes an interpretable and calibrated Artificial Intelligence framework for multiclass banana disease-pattern characterization based on radiomic feature analysis of RGB leaf images.
SoybeanNeRF / 3D Gaussian SplattingRGB / grayscaleFruitSegmentation
Neural Radiance Fields (NeRF) have been widely adopted for reconstructing high-quality 3D scenes from 2D RGB images. However, achieving accurate 3D object segmentation within these reconstructed scenes remains challenging. Existing NeRF-based segmentation methods either rely on post-processing (SA3D), which produces noisy point clouds due to the absence of density field optimization, or employ joint training with additional segmentation heads (FruitNeRF), which can lead to suboptimal performance due to conflicting learning objectives. In this work, we propose InvNeRF-Seg (Input-substitution NeRF for Segmentation), a two-stage fine-tuning strategy for 3D object segmentation that preserves the original NeRF architecture and loss function entirely. We first train a standard NeRF on RGB images and then fine-tune it using 2D segmentation masks formatted as RGB-like inputs, without introducing any architectural modifications or additional loss functions. This input-substitution approach reshapes the density field to align with object regions while suppressing background density. We validate InvNeRF-Seg through comprehensive ablation studies examining the roles of density and color MLPs, loss function choices, and training strategies. Field density analysis reveals consistent semantic refinement: densities of object regions increase while background densities are suppressed. Experiments on synthetic fruit datasets and real-world soybean imagery demonstrate that InvNeRF-Seg produces cleaner 3D segmented point clouds compared to both SA3D and FruitNeRF, enabling more accurate downstream object counting. The method is further validated on a self-collected soybean dataset to demonstrate its applicability in real-world agricultural scenarios. Our code is available at https://github.com/ZJiangsan/InvNeRF-Seg .
Why it matches plant phenotyping methods植物画像から3D物体領域を抽出するNeRFベース手法の開発・比較検証が中心で、果実・ダイズ画像を対象に物体カウントへ応用しているため、植物器官の形態・数量推定に関わるフェノタイピング手法として含める。
abstractIn this work, we propose InvNeRF-Seg (Input-substitution NeRF for Segmentation), a two-stage fine-tuning strategy for 3D object segmentation
This dataset contains images of Manalagi apple diseases from Indonesia. Data collection was conducted from August 2024 to June 2026. Data were collected in apple orchards. All images were captured under natural environmental conditions. A total of 1168 unique Manalagi apple specimens were successfully documented. The specimens consisted of both healthy and diseased fruit. This dataset comprises four classes: Healthy, Anthracnose, Black Pox, and Powdery Mildew. Each specimen was observed and photographed directly. The documentation process yielded approximately 5100 raw images. The images were captured using various smartphone cameras and DSLR cameras. Each device has different camera specifications. The image size depends on the device used. Images that passed quality inspection were selected for the next stage. Each fruit specimen is cropped from the selected raw image. Each image was then labeled according to its disease class. The image size was standardized to 1024 × 1024 pixels. All images were saved in JPEG format. The curation process yielded 482 images. Each image represents a distinct fruit specimen.
Why it matches plant phenotyping methodsリンゴ果実の健全・病害状態を画像で記録し、分類用データセットとして構築・キュレーションした研究であり、植物病害表現型の取得方法と再利用可能なデータセットが中心です。
titleImage dataset of manalagi apple fruits for multi-class disease classification using deep learning.
Reproduction assets foundThe paper is a Data in Brief article describing a public Mendeley Data repository of Manalagi apple fruit disease images (raw, curated, and augmented), directly usable for plant disease phenotyping/classification.Dataset · publicRepository name: Mendeley Data
Data identification number: DOI: 10.17632/9zgkwwv9j8.6
Direct URL to data: https://data.mendeley.com/datasets/9zgkwwv9j8/6Open asset ↗Mendeley Data · 10.17632/9zgkwwv9j8.6html-lines:97-124Plant phenotyping relevance match · UnverifiedCrossref · checked 14 Sept 2026
Potato leaf disease is one of the main problems in potato cultivation because it can reduce plant quality, decrease crop yield, and cause economic losses for farmers. Manual disease detection still has limitations because it depends on farmers’ experience and is prone to errors, especially when disease symptoms have similar visual characteristics. This study aims to apply the Convolutional Neural Network (CNN) algorithm to predict potato leaf diseases based on digital images. The dataset used in this study was obtained from Kaggle and consisted of 1,500 potato leaf images divided into three classes: healthy leaves, early blight, and late blight. The research stages included dataset collection, data splitting into training, testing, and validation data, CNN modeling using Jupyter Notebook, model training with 50 epochs, model evaluation using a Confusion Matrix, and model implementation into a web-based system using Flask. The test results show that the CNN model was able to classify potato leaf diseases with an accuracy of 97%. These results indicate that CNN is effective in recognizing visual patterns in potato leaf images, such as color changes, spots, and leaf damage. This study is expected to serve as a basis for developing an early detection system for potato leaf diseases that is faster, more accurate, and easier for farmers to use.
Why it matches plant phenotyping methodsジャガイモ葉の画像から病害状態を推定するCNN手法が研究の中心であり、モデル評価と実装も行っているため、植物フェノタイピング手法として含める。
abstractThis study aims to apply the Convolutional Neural Network (CNN) algorithm to predict potato leaf diseases based on digital images.
Perilla ( Perilla frutescens ) is an important oilseed crop in East Asia with high nutritional and economic value. Seed size and seed coat color are key agronomic traits influencing yield, oil quality, and market preference. However, genetic studies in perilla remain limited by small population sizes and low-throughput phenotyping, restricting their application in breeding. A large-scale genome-wide association study (GWAS) was conducted by integrating high-throughput image-based phenotyping in a panel of 493 perilla accessions. Measurements of seed morphology, including area, perimeter, and length, and color traits (RGB components) were obtained. Genotyping-by-sequencing generated high-quality single-nucleotide polymorphism (SNP) datasets, and GWAS was performed using four statistical models (general linear model, mixed linear model, FarmCPU, and BLINK). A total of 44 significant trait-SNP associations were identified, corresponding to 20 unique SNPs, as several SNPs (including those on chromosomes 6, 10, 15, and 17) were associated with multiple correlated traits. Linkage disequilibrium-based analysis showed candidate genes involved in phenylpropanoid metabolism and carbohydrate pathways. Predicted protein-altering variants, including non-synonymous and stop-gained mutations, were detected in key genes. Derived cleaved amplified polymorphic sequence markers developed near peak SNPs distinguished phenotypic differences between allelic groups, demonstrating their effectiveness for trait differentiation. This study represents the first large-scale GWAS integrating image-based phenotyping for seed traits in perilla and provides candidate dCAPS markers with potential applicability to marker-assisted selection, pending validation in independent breeding populations. These findings offer valuable genetic resources and a practical framework for molecular breeding and crop improvement in perilla.
Why it matches plant phenotyping methods大規模GWASに統合された高スループット画像ベース表現型解析が中心で、種子形態・色形質の抽出方法を実質的に適用しているため。
abstractA large-scale genome-wide association study (GWAS) was conducted by integrating high-throughput image-based phenotyping in a panel of 493 perilla accessions.
Precision agriculture increasingly relies on detailed structural information, such as canopy height and canopy volume, to enhance crop health monitoring and operational safety. However, existing methods based on costly LiDAR or RGB-D sensors are often impractical for large-scale deployment in dynamic and unstructured horticultural environments. Furthermore, conventional 2D segmentation and SLAM-based pipelines typically generate sparse, geometrically inconsistent semantic maps which are insufficient for actionable structural analysis in agricultural applications. To overcome these limitations, we propose a monocular 3D structural mapping framework tailored for horticultural plants via semantic scene completion. At inference, the proposed model takes a single RGB image as input and predicts voxel-wise geometry and semantics, from which task-oriented structural maps, including canopy height, canopy volume, and obstacle-aware traversability layers, are derived. Specifically, we first introduce a Depth-Aware Decoder Module that explicitly recovers depth in the spatial domain and fuses 2D-to-3D features, thereby mitigating depth ambiguity and reducing reliance on accurate pose. Second, an NCS-Guided Geometry Encoder is designed to inject normalized depth into voxel positional embeddings, enabling self-attention to perform global relational modeling within a depth-aware geometric coordinate system. In addition, a Global Encoder is utilized to refine local structural details, while an occupancy head produces the final 3D semantic completion outputs. We construct a horticultural 3D semantic scene dataset using an RGB-D sensor, which serves as a benchmark for evaluating our method, while the deployed model remains RGB-only. Extensive quantitative and qualitative experiments are conducted on both the Semantic-KITTI dataset and our dataset. On our dataset, the method achieves 82.31% occupancy IoU, 84.26% mIoU, and 86.25% precision. Beyond voxel-level evaluation, manual field measurements further show canopy height MAE values of 0.019-0.026 m and canopy volume proxy relative errors of 8.4%-11.4%. These results demonstrate the effectiveness of our approach in real-world agricultural scenarios, providing actionable structural insights for crop monitoring and autonomous robotic operations.
Why it matches plant phenotyping methods単眼RGB画像から植物の樹冠高・樹冠体積などの構造形質を推定する3Dフェノタイピング手法を開発し、データセット構築と実測検証も行っているため、方法が研究の中心である。
abstractwe propose a monocular 3D structural mapping framework tailored for horticultural plants via semantic scene completion.
Accurate color representation is critical for UAV-based crop phenotyping, yet UAV images are often distorted by variable illumination and camera exposure settings. Here, we propose Lite U-net FiLM (LUF-net), a lightweight U-net framework integrated with feature-wise linear modulation (FiLM) layers, which incorporates both multispectral-derived irradiance and camera exposure parameters as auxiliary inputs. This metadata-aware design enables dynamic modulation of intermediate image features, allowing the network to disentangle unified canopy color from environmental artifacts. Experiments conducted across diverse crop types (soybean, rice, maize), flight altitudes (6m, 12m, 25m, 40m), and illumination conditions (overcast skies, cloudy, sunny, early morning) demonstrate that the LUF-net model substantially reduces the mean absolute percentage error to below 5.5%, outperforming conventional Gray-world, U-net, AlexNet-MLP, and SIDBlock-MLP methods. Ablation experiments further show that both irradiance and exposure metadata provide complementary information, while FiLM-based conditioning effectively integrates these acquisition parameters into feature learning, jointly contributing to improved color reconstruction performance. Moreover, correlation analysis indicates that the model's color reconstruction errors are weakly dependent on irradiance and exposure settings, confirming that LUF-net reduces sensitivity to external imaging conditions while maintaining physically meaningful and robust corrections.
Why it matches plant phenotyping methodsUAV画像の色校正手法を開発し、複数作物・撮影条件で性能検証しており、作物フェノタイピングの画像取得・補正が中心である。
abstractAccurate color representation is critical for UAV-based crop phenotyping, yet UAV images are often distorted by variable illumination and camera exposure settings.
Reproduction assets foundThe paper's data and code availability statement explicitly points to a public GitHub repository containing training/evaluation/inference code, pretrained model weights, and example data for the LUF-net color calibration method.Code · publicThe data and source code for model training, evaluation, and inference, together with pretrained model weights, example data, and detailed usage instructions, is publicly available at: https://github.com/wangchufeng3652/color-correction.Open asset ↗wangchufeng3652/color-correctionhtml-lines:278-299Code / dataset availability confirmedEurope PMC · checked 14 Sept 2026
Farmers must be able to estimate their crop yields at various growth stages for effective management of their farms and to enable them to interact with cooperatives or traders as early as possible. Here we developed an AI-based cocoa pod detection method using low resolution colour images of cocoa trees on farms in Côte d'Ivoire. We compared nano and extra-large architectures of six neural networks, including Faster RCNN, Baidu's Real-Time Detection Transformer (RTDetr), Detr-ResNet Vision Transformer (ViT), YOLOv5, YOLOv8 and YOLOv11. These networks were trained with 7,850 annotated cocoa pods on 400 low resolution images, and validated in two independent datasets: a 42 low resolution images containing 990 annotated pods, and a 100 low resolution images containing 2,400 annotated pods. The performances of the nano YOLOv8 and YOLOv11 networks were 2% higher than that of the RTDetr networks and 5% higher than that of the YOLOv5, ViT and Faster RCNN networks with an F1-score of 77% on all images and up to 90% on foreground trees. The dominance of nano architectures suggests that the extra-large architectures, which contain 20-30-times more neurons, may not have been fully trained. The study of learning performance curves showed that extra-large networks were unable to outperform nano networks, which contradicts the theory. After review, the annotated dataset was found to contain inconsistencies. The inconsistency of the training and validation data and their limited quantity restricted the objectivity of comparisons between network architectures. Finally, although the average detection performance of RTDetr for cocoa pods was only 2% lower than that of the YOLOv8 network, it was definitively excluded from the candidate models because its per-image processing time was 15-20% higher than that of YOLOv8 and YOLOv11. However, with a performance sensitivity to data of less than 0.5%, YOLOv8 Nano became the best option.
Why it matches plant phenotyping methodsカカオ果実を画像から検出・定量するAI手法を開発し、複数モデルと独立データセットで性能比較・検証しており、植物フェノタイピング手法が中心である。
abstractHere we developed an AI-based cocoa pod detection method using low resolution colour images of cocoa trees on farms in Côte d'Ivoire.
Reproduction assets found保存済みの本文根拠を更新済みルールで再検証し、公開資産1件を確認しました。Dataset · publicData Availability: The data used in the study can be downloaded from CIRAD’s data verse at https://doi.org/10.18167/DVN1/8COJBB .Open asset ↗10.18167/DVN1/8COJBBlines:129-140Plant phenotyping relevance match · UnverifiedEurope PMC · Crossref · checked 5 Sept 2026
Abstract Outbreaks of plant diseases are major threats to world food security particularly in areas where real-time monitoring and quick decision support are constrained by low-power edge gadgets and untrustworthy connectivity. In order to overcome these issues, this paper presents a FPGA-Accelerated IoT implementation of a Causal-Attention Multi-Modal Deep Learning Network, named EpiFusionNet-Edge, that can be applied to monitor crop diseases with real-world farming scenarios with high precision and scalability. The framework incorporates five data modalities that are complementary in nature and they include RGB leaf pictures, microscopic foldscope images, UAV hyperspectral signatures, microclimate IoT sensor measurements and region-specific pathogen/pest pressure indexes giving a complete picture of the health of the plant. Dual causal-attention mechanism is proposed to simulate both spatial and temporal environmental factor activation, which helps to detect and make predictions at the early stage and provides an explanatory logic behind the decisions. Multi-task learning enables classification of diseases, quantification of their intensity at the level of a micro-prediction and prediction of outbreaks in the short term (1–30 days). In order to achieve deployability in resource-constrained settings, the proposed deep learning architecture is ensemble-distilled, structurally pruned, and INT8-quantized, and hardened on a Xilinx Zynq-7000 FPGA platform. The FPGA accelerator is 43.2x faster inference, 88 percent less power usage, and less than 10 ms latency, which allows real-time execution of continuous field monitoring with IoT sensors. Cross-condition assessment on multi-domain datasets shows that there are great improvements on cross-environment generalization rates with 98.6% classification accuracy, 92.7% severity estimation accuracy and less than 3.5% degradation with domain shift. Grad-CAM + + and causal feature traceability further add interpretability with the focus of the model and the pathological indicators proven by experts. The findings show the promise of using a combination of IoT sensing, multi-modal AI fusion, and FPGA hardware acceleration to develop a deployable and scalable and transparent system with regard to precision agriculture. This paper creates a roadmap to a new generation of smart farming systems that are able to conduct disease surveillance and actively protect crops at the periphery in an autonomous manner.
Why it matches plant phenotyping methods植物病害の画像・センサー観測から病害強度を定量化するマルチモーダル・エッジ推論基盤を開発し、精度・速度・消費電力・ドメインシフトを評価しているため、植物フェノタイピング手法が中心である。
abstractthis paper presents a FPGA-Accelerated IoT implementation of a Causal-Attention Multi-Modal Deep Learning Network, named EpiFusionNet-Edge
ABSTRACT Banana diseases impose severe production losses in tropical smallholder farming systems, yet accurate in-field visual diagnosis remains difficult: symptom expression varies across cultivars and growth stages, and several diseases produce morphologically overlapping foliar signs. We developed a probabilistic image-recognition framework for detecting five economically important banana diseases — Xanthomonas Wilt, Banana Bunchy Top Disease, Fusarium Wilt (Panama disease), Yellow Sigatoka, and Black Sigatoka — from in-field photographs, without any disease-specific fine-tuning of the vision backbone. The approach extracts frozen 1,152-dimensional embeddings from the DINOv3 vision foundation model and couples them with a conditional normalizing flow, trained on four publicly available datasets spanning diseased banana plants, healthy tissue, non-banana vegetation, and general natural imagery. On an independent test set the model achieved F1 scores exceeding 0.98, average precision values of 0.968–0.999, and AUROC values of 0.997–1.000 across all five diseases evaluated as binary detection problems. Multi-class accuracy was near-perfect, with limited confusion between Yellow Sigatoka and Black Sigatoka — a biologically plausible ambiguity attributable to overlapping early-infection foliar symptoms. Because the normalizing flow estimates explicit conditional probability densities rather than decision boundaries, two complementary log-likelihood ratios can be derived: a disease ratio comparing each disease class against healthy banana, and a plant ratio comparing banana against non-banana imagery. Together these define an interpretable two-dimensional diagnostic space that simultaneously quantifies evidence for disease presence and image relevance, cleanly separating diseased plants, healthy plants, and out-of-distribution images while flagging uncertain predictions for confirmatory testing. Inference on frozen embeddings is lightweight and compatible with smartphone deployment, providing a scalable, uncertainty-aware diagnostic tool for smallholder farming systems and disease surveillance programmes.
Why it matches plant phenotyping methodsバナナ葉の病徴を圃場画像から直接推定する確率的画像認識手法を開発し、独立テストセットで性能検証しているため、植物病害状態のフェノタイピング手法が中心である。
abstractWe developed a probabilistic image-recognition framework for detecting five economically important banana diseases
Smartphone-based precision spraying aims to restrict chemical application to target organs, but it also requires an organ-level perception model that operates on RGB alone at deployment, without dedicated depth sensors. Combining with monocular depth foundation models such as Depth Anything V2, RGB images can be enriched with dense pseudo-depth during training, turning close-range crop segmentation into a privileged-depth-guided learning problem where depth is available at training but absent at inference. Based on such close-range handheld capture, intra-foreground depth separability collapses while only the foreground–background depth gap remains partially preserved, exposing an imaging regime that we refer to as Close-Range Depth Degeneracy (CRDD) and that undermines the discriminative value of monocular depth for organ segmentation. Even with the available privileged-depth signal, existing RGB-D fusion and depth-branch hallucination routes still either introduce inference-time depth dependence or fail to retain the depth-conditioned effect inside RGB-side parameters under CRDD. How monocular pseudo-depth can serve as training-only privileged information for RGB-only close-range crop spraying under CRDD remains insufficiently studied. To this end, this paper proposes a privileged-depth modulation framework, MDM-Seg, by converting monocular depth from an inference-time modality into a CRDD-conditioned training signal to remove the inference-time depth dependence. Specifically, we characterize CRDD and propose a four-indicator depth-usability diagnostic that delivers a pre-training verdict on the target dataset, with verdicts validated against post-training segmentation outcomes. Then, we formalize the CRDD-conditioned training intervention as a depth-conditioned affine network (DepAN) that injects depth-gradient cues into decoder features, and define a depth hard-pixel mining loss (DHPM) to reweight per-pixel cross-entropy with depth-value evidence, both operating inside the main segmentation stream. Finally, we deploy MDM-Seg on RGB input alone and derive a lightweight consistency trace between the RGB-only and depth-conditioned heads, producing an image-level reliability cue for targeted-versus-fallback spraying without depth at inference. Extensive experiments on Pepper-Field-3940 demonstrate that our MDM-Seg achieves effective depth-conditioned RGB-only segmentation under CRDD, i.e., depth-conditioned modulation retained inside RGB-side parameters with no inference-time depth dependence, such as mIoU 0.8995 with the ResNet-101 configuration ( + 3.73 pp over the RGB-only DeepLabV3+ baseline), a lightweight training-side DepAN module, and an image-level reliability cue rank-correlated with segmentation quality (Spearman ρ = 0.74 ). On a Jetson Orin NX 16 GB with TensorRT FP16, the complete deployment-oriented MobileNetV2 dual-head prescription pipeline runs at 23.87 ms per image, corresponding to 41.89 FPS, 0.62 GB peak memory, and 0.56 J per image.
Why it matches plant phenotyping methods植物器官の画像セグメンテーション手法を開発・評価し、散布対象器官の状態をRGB画像から推定することが中心であるため。
abstractthis paper proposes a privileged-depth modulation framework, MDM-Seg
Unmanned aerial vehicle (UAV) photogrammetry offers a cost-effective approach to tree-level detection, however, Structure-from-Motion (SfM) outputs are sensitive to processing choices and site conditions, which can alter canopy representation and reduce individual-tree detection accuracy. Here, we systematically evaluate how SfM reconstruction quality and depth-map filtering influence RGB-only individual-tree detection under controlled acquisition conditions. Objectives were to (i) identify an optimal SfM-derived point-cloud configuration for delineating individual trees, and (ii) implement and test a segmentation workflow (local-maxima treetop detection plus Dalponte2016 in lidR) for detecting and counting trees. We assessed RGB-only SfM for individual-tree detection (ITD) across thirteen 1.21-ha loblolly pine ( Pinus taeda ) plots located in two counties in the state of Alabama in the southeastern United States; eight even-aged plantations and five mixed pine-hardwood stands, while holding image acquisition parameters constant. Using Agisoft Metashape Professional (Agisoft LLC, St. Petersburg, Russia), dense-cloud quality (Lowest, Low, Medium, High, Ultra High) and depth-map filtering (Disabled, Mild, Moderate, Aggressive) were varied in a 5 × 4 full-factorial design; assessment metrics included point-cloud density, canopy-surface completeness, canopy-height-model (CHM) agreement with field heights, and ITD precision/recall/F1. We identified a single high-resolution configuration (Ultra High + Disabled) by screening parameter sets for structural accuracy and suppression of false peaks. Using this configuration, CHMs matched field heights in Washington County, Alabama (R 2 = 0.96; RMSE = 0.44 m; bias = − 0.01 m) and in Cullman County, Alabama (R 2 = 0.44; RMSE = 1.14 m; bias = − 0.09 m); pooled performance was R 2 = 0.98; RMSE = 0.54 m; bias = − 0.01 m. ITD accuracy at the primary 3 m match radius yielded a precision of 0.03; recall = 0.29; F1 = 0.05 in the even-aged plantations (Washington) and a precision of 0.03; recall = 0.12; F1 = 0.05 in mixed pine–hardwood stands (Cullman); pooled F1 = 0.05. The selected parameters and workflow are reproducible and transferable, provide insight into RGB-SfM ITD performance, and indicate when lidar remains preferable for crown delineation.
Why it matches plant phenotyping methodsRGB-SfMによる個体樹の検出・樹高推定と、SfM設定およびセグメンテーションワークフローの系統的評価が研究の中心であり、植物の樹冠構造・樹高という形態形質を抽出する方法を検証している。
abstractwe systematically evaluate how SfM reconstruction quality and depth-map filtering influence RGB-only individual-tree detection
Reproduction assets foundThe paper's Code availability statement deposits the authors' SfM/ITD processing scripts publicly on OSF (DOI 10.17605/OSF.IO/UXBCZ). Phenotype/field datasets are only available on request, so they are not public assets.Code · publicThe workflow and processing scripts used in this study are publicly available through the Open Science Framework
(OSF) repository: Singh and Narine, [32]. Code Repository for Optimizing SfM Parameters for RGB-Only Individual-Tree
Detection in Loblolly Pine and Mixed Pine-Hardwood Stands. https://doi.org/10.17605/OSF.IO/UXBCZ.Open asset ↗10.17605/OSF.IO/UXBCZpdf-page:12 lines:1-70Code / dataset availability confirmedEurope PMC · checked 5 Sept 2026
Plant diseases significantly affect agricultural productivity and global food security, while accurate disease identification remains challenging because of uncertain and overlapping visual symptoms in leaf images. Existing deep learning approaches often require large annotated datasets and suffer from limited interpretability in practical agricultural environments. This study presents a multi-parameter improved fuzzy soft set-based framework for plant disease classification using tomato leaf images from the PlantVillage dataset. The objective is to develop an interpretable and reliable classification model capable of handling uncertainty in plant disease patterns through feature-driven fuzzy similarity analysis. The methodology integrates image preprocessing, color and texture feature extraction, variance-based feature weighting, prototype generation using K-means clustering, and fuzzy similarity computation using Mahalanobis distance and Gaussian membership functions. RGB, HSV, and Gray-Level Co-occurrence Matrix (GLCM) features are extracted from standardized leaf images and evaluated within an improved fuzzy soft classification framework. Performance comparison is carried out using machine learning models including Support Vector Machine (SVM), Random Forest (RF), Linear Discriminant Analysis (LDA), and Naive Bayes (NB) implemented in Python using Scikit-learn libraries. Experimental simulation results demonstrate that the proposed framework achieves competitive classification performance while preserving interpretability and robustness under uncertain feature distributions. Performance evaluation is conducted through accuracy analysis, ROC-AUC curves, confusion matrices, ablation studies, and Wilcoxon Signed-Rank statistical testing. The proposed Improved Fuzzy Soft model achieved an accuracy of 88.57% which is less than LDA (94.92%), Random Forest (97.78%) and SVM (97.94%) classifiers. However, in the cross data set validation, the proposed Improved Fuzzy Soft model achieved an accuracy of 67.35% which is greater than LDA (51.02%), Random Forest (51.02%) and SVM (55.10%) classifiers. Statistical validation using the Wilcoxon Signed-Rank Test produced a p-value of [Formula: see text], confirming that the performance difference between the Improved Fuzzy Soft framework and the Random Forest classifier is statistically significant under the current experimental setting.
Why it matches plant phenotyping methodsトマト葉画像から植物病害状態を推定する解釈可能な画像解析・分類フレームワークを開発し、複数モデル、交差データセット検証、アブレーション、統計検定で評価しており、病害表現型の取得・抽出手法が中心である。
abstractThis study presents a multi-parameter improved fuzzy soft set-based framework for plant disease classification using tomato leaf images from the PlantVillage dataset.
Reproduction assets foundThe paper uses public tomato leaf image datasets (PlantVillage and PlantDoc from Kaggle) as phenotyping inputs and states the authors' Improved Fuzzy Soft Framework implementation is publicly available on Zenodo with source code and reproduction instructions.Dataset · publicThe dataset analyzed during the current study are available in the repository:
https://www.kaggle.com/datasets/abdallahalidev/plantvillage-datasetOpen asset ↗kaggle.com/datasets/abdallahalidev/plantvillage-datasetpdf-page:24 lines:1-75Code · publicThe implementation of the proposed Improved Fuzzy Soft Framework is publicly available through the Zenodo repository:
https://doi.org/10.5281/zenodo.20570546
The repository contains the source code, documentation, and instructions required to reproduce the experiments reported in
this study.Open asset ↗zenodo · 10.5281/zenodo.20570546pdf-page:25 lines:1-74Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · checked 15 Sept 2026
Abstract Detecting plant leaf diseases at an early stage is one of the most important requirements for sustainable agriculture, increasing crop productivity, and achieving the global Sustainable Development Goals (SDGs). However, accurately recognizing them in real-world farm fields can still be difficult due to factors such as background complexity, changes in light conditions, and very similar looking classes from a visual standpoint. In order to solve these problems, the authors here present a new Multi-Scale Feature Fusion (MSFF) model that can offer robust and highly accurate performance in identifying plant leaf diseases. Firstly, the brand new hybrid method starts with a U-Net segmentation designed exclusively to separate the diseased parts and thus allow the classification to be more robust. Next, Rank Order Fuzzy (ROF) is implemented to get rid of the background while still maintaining the edges, and the additional data is used for the network to generalize better. The color distribution is then analyzed to determine the variations in color brought about by the infection. In terms of features, EfficientNet is paired with an Attention-based Autoencoder to produce both spatially detailed global features and compact latent representations. The two sets of features are then combined through Canonical Correlation Analysis (CCA) which not only identifies the dependencies between the features but also enhances the discriminative strength. The final fused feature set is fed to module based on YOLO for detection and classification in order to obtain the final result of the plant disease identification system which is both accurate and fast. Among other datasets, the model has been tested on different apple leaf datasets such as FGVC7, AppleLeafSet, PlantVillage Apple, Kaggle Apple Leaves, and ATLDSD, which contain five disease classes. The results of the experiments indicate a classification accuracy of 99.95%, thus the model is superior to several state-of-the-art deep learning and classical machine learning methods. Statistical methods like five-fold cross-validation, paired t-tests ( p
Why it matches plant phenotyping methods植物葉の病変部位を画像から自動抽出し、病害の識別・分類を行う画像ベースのフェノタイピング手法を開発・評価しており、手法が研究の中心である。
abstractthe authors here present a new Multi-Scale Feature Fusion (MSFF) model that can offer robust and highly accurate performance in identifying plant leaf diseases
Grapevine yellows diseases, including Flavescence Dorée (FD) and Bois Noir (BN), severely affect vineyards, making reliable detection methods essential. Detecting symptoms, particularly in Chardonnay variety, remains challenging due to year-to-year vine variability and the presence of symptoms that can be misleading. We present a dataset of multispectral images of grapevine leaves collected over four years (2021–2024). The dataset includes healthy leaves, grapevine yellows - overwhelmingly BN, and three other diseases (Esca, Discoloration, and Leafroll) with visually similar symptoms particularly for Leafroll, providing a diverse benchmark for disease recognition. Ground-truth labels were assigned based on field inspections conducted by Comité Champagne expert viticulturists who selected the vines before acquisitions and analyzed the images after acquisition. Images were acquired under laboratory conditions using a DJI Phantom 4 Multispectral camera from leaves collected in the Plumecoq Comité Champagne experimental vineyards. Leaves were collected on designated vines and placed on polystyrene boards on a 2x2 grid. One RGB image and five multispectral images corresponding to the Blue (B - 450 nm ± 16 nm), Green (G - 560 nm ± 16 nm), Red (R - 650 nm ± 16 nm), Red-Edge (RE - 730 nm ± 16 nm), and Near-Infrared (NIR - 840 nm ± 26 nm) bands were captured using the same acquisition system (camera-to-board distance of approximately 90 cm) over the four years.. For the vast majority, ambient light was used; for some acquisitions, indirect halogen spotlights were used. The 512x512 leaf images were manually cropped from the acquired original images with the same offsets between the bands; offsets were already present due to the parallax phenomenon. The dataset is organized into six folders (B, G, R, RE, NIR, and RGB). Each leaf is represented by one RGB image and five corresponding multispectral band images, enabling multimodal analysis and machine learning for grapevine disease detection. Each image filename encodes acquisition and annotation information as follows: Characters 1–2: acquisition year (21 = 2021, 22 = 2022, 23 = 2023, 24 = 2024). Character 3: image type/band (0 = RGB, 1 = B, 2 = G, 3 = R, 4 = RE, 5 = NIR). Character 4: class (J = Grapevine Yellows, T = Healthy, S = Esca, E = Leafroll, D = Discoloration). Character 5: illumination condition (0 = ambient light, G = additional left halogen spotlight, D = additional right halogen spotlight, 2 = additional both left and right halogen spotlightsl). Characters 6–9: image identifier, numbered sequentially from 0000. Character 10: relabelling status. After image acquisition, all leaves were independently reviewed by expert viticulturists on the recorded RGB images. A value of 1 indicates that the expert confirmed the original field label, 0 indicates that the original label was considered incorrect based on the information visible in the RGB image, and D indicates that the sample was reclassified as Discoloration. This dataset complements the "Multi-annual spectral data of Chardonnay grapevine leaves" dataset published on Recherche Data Gouv. While the previous dataset contains spectral measurements, the present dataset provides multispectral images acquired from the same grapevine leaves, enabling multimodal analyses that combine spectral signatures with image-based information. Note, however, that there is no bijection between the respective files.
Why it matches plant phenotyping methodsブドウ葉の病徴をマルチスペクトル画像で取得し、専門家ラベル付きの疾病認識用データセット／ベンチマークとして提供しており、植物状態の画像ベース表現型計測が中心である。
abstractWe present a dataset of multispectral images of grapevine leaves collected over four years (2021–2024).
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 5 Sept 2026
Fall dormancy (FD) and forage yield (FY) are two key traits in alfalfa ( Medicago sativa L.) breeding programs. However, genetic progress has remained limited over the past decades, largely due to the complexity of alfalfa breeding and the reliance on labor-intensive phenotyping methods. High-throughput phenotyping (HTP) using unmanned aerial vehicles (UAVs) represents a promising alternative for rapid and non-destructive crop evaluation. The objectives of this study were to i) estimate FD using UAV-derived canopy height and RGB vegetation indices and ii) evaluate the predictive performance of machine learning (ML) models for FY estimation. A total of 210 alfalfa populations with diverse genetic backgrounds were evaluated over two growing seasons (2023 to 2025) across seven harvests under Mediterranean conditions in central Chile. FY was measured manually, while FD was estimated using both manual and UAV-based approaches. A total of 19 RGB-derived indices (VIs) including plant height (PH) were extracted and used as predictor variables. Five complex predictive ML models were evaluated: PLS, PCR, SVM, ANN, and MLR. The results showed that UAV-derived FD was significantly correlated with FD obtained through conventional methods ( R 2 = 0.88). The automated UAV-based FD phenotyping framework demonstrated slightly higher precision ( R 2 = 0.92) and broad-sense heritability ( H 2 = 0.69) compared to manual measurements ( R 2 = 0.87–0.89; H 2 = 0.64), providing a more reliable selection tool for breeders. Among the tested ML models, SVM and ANN achieved the highest accuracy ( R 2 ≈ 0.73) for FY prediction. These findings demonstrate that integrating low-cost RGB imagery with complex modeling offers a promising avenue that could assist in refining future selection strategies for this genetically complex species.
Why it matches plant phenotyping methodsUAV画像からアルファルファの休眠性と収量関連形質を推定する高スループット表現型解析手法を開発・検証し、手動測定との比較と機械学習モデル評価を行っているため、方法が研究の中心である。
abstractHigh-throughput phenotyping (HTP) using unmanned aerial vehicles (UAVs) represents a promising alternative for rapid and non-destructive crop evaluation.
Early crop establishment strongly influences plant performance and yield, making seedling emergence an important trait in crop phenotyping, breeding, and stress physiology studies. However, emergence monitoring is still commonly performed manually and typically records only the final emergence percentage, limiting the analysis to other dynamic observations. Automated image-based approaches are promising but remain challenging due to the small size of plant structures, heterogeneous soil backgrounds, and variability across imaging systems. Here, we present SPROUT (AI- based S eedling PR edicti O n and trait extraction U sing RGB T ime-series), a low-cost RGB imaging pipeline for automated prediction of crop emergence dynamics and trait extraction. The system integrates instance segmentation, object-detection–based data reduction, and temporal deep learning to estimate the emergence time of individual seedlings from RGB image sequences. The pipeline then automatically reconstructs emergence curves and extracts associated traits, including final emergence percentage, EC50, and emergence synchronicity. SPROUT was developed and evaluated using barley and wheat datasets acquired with different RGB cameras under controlled growth-chamber conditions. In the development and retraining settings, the best-performing TCN model achieved 90.0% per-well accuracy with a ± 2h tolerance, supporting accurate emergence curve reconstruction. In an independent inference-only dataset, the model still captured approximate emergence dynamics, although accuracy decreased to 59.3%, indicating that SPROUT is best used as a modular pipeline that can be retrained or fine-tuned for new crop, camera, or experimental domains. A cadmium-stress case study in two contrasting wheat genotypes showed that SPROUT-derived traits captured genotype-specific establishment strategies associated with growth and metabolic responses.
Why it matches plant phenotyping methodsRGB時系列画像から出芽動態を推定し、出芽率・EC50・同時性などの形質を抽出するパイプラインを開発・評価しており、植物フェノタイピング手法が中心である。
abstractHere, we present SPROUT (AI- based S eedling PR edicti O n and trait extraction U sing RGB T ime-series), a low-cost RGB imaging pipeline for automated prediction of crop emergence dynamics and trait extraction.
Reproduction assets foundThe paper's SPROUT emergence-prediction pipeline code is publicly available on GitHub with explicit availability language, and raw images plus morphology/metabolic data are deposited on Zenodo (10.5281/zenodo.18889863). The GitHub URL is in allowed_urls; the Zenodo DOI is not, so only the code asset is listed as an ad-Code · publiccan be found online at https://doi.org/10.1016/j.compag.2026.112184.Data availability
The raw images and raw data for the morphology and metabolic
profiling on the case study are available in ZENODO (10.5281/zen
odo.18889863), and the code for the machine learning pipeline and
emergence curve analysis are available on GitHub (https://github.com/kit-pef-czu-cz/sprout-emergence-prediction).References
Albarenque, S., Basso, B., Davidson, O., Maestrini, B., Melchiori, R., 2023. Plant
emergence and maize (Zea mays L.) yield across multiple farmers’ fields. Field Crops
Res. 302. https://doi.org/10.1016/j.fcr.2023.109090.Arsovski, A.A., Galstyan, A., Guseman, J.M., Nemhauser, J.L., 2012.
PhotomorpOpen asset ↗kit-pef-czu-cz/sprout-emergence-predictionpdf-raw-page:13 lines:78-112Code / dataset availability confirmedOpenAlex · Europe PMC · checked 5 Sept 2026
Orthoimagery derived from unmanned aerial vehicles (UAVs) has become a valuable data source for crop-growth monitoring. Individual plant-level (IPL) information enables high-throughput analyses by capturing plant-to-plant variability within fields. However, reliable IPL-based analysis requires accurate extraction of plant-specific regions, which remains challenging in soybean cultivation due to weed interference and canopy overlap. This study proposed an automatic preprocessing framework for IPL soybean growth monitoring that integrates deep-learning-based semantic segmentation with a furrow-guided region of interest (ROI) generation strategy using UAV imagery. A segmentation model was developed using combinations of RGB and multispectral orthoimagery, and a furrow line detection algorithm was designed to generate IPL ROIs aligned with crop rows. The ensemble model combining U-Net, DeepLabV3+, and SegFormer achieved the most stable performance (F1-score up to 0.94 and IoU up to 0.89). The furrow-guided ROI generation algorithm also accurately estimated crop counts, showing strong agreement with manual observations (R² = 0.90 and RMSE = 6.35). The generated IPL ROIs enabled accurate quantification of growth-related features, with strong agreement between automatically generated and manually delineated ROIs (R² > 0.90). Overall, the proposed preprocessing framework provides a practical and scalable solution for UAV-based high-throughput phenotyping in soybean and other ridge-based cropping systems.
Why it matches plant phenotyping methodsUAV画像から個体単位の植物領域を抽出し、成長形質を定量化する前処理・セグメンテーション手法が研究の中心であるため。
abstractThis study proposed an automatic preprocessing framework for IPL soybean growth monitoring that integrates deep-learning-based semantic segmentation with a furrow-guided region of interest (ROI) generation strategy using UAV imagery.
Reproduction assets foundThe authors state that the complete implementation of their IPL soybean preprocessing pipeline (semantic segmentation + furrow line detection) is publicly available on Zenodo. The annotated sample dataset, however, is only available upon request from the corresponding author, so it is not a public asset.Code · publicThe complete implementation of
this pipeline is publicly available at https://doi.org/10.5281/zenodo.21095307.Open asset ↗zenodo · 10.5281/zenodo.21095307pdf-page:7 lines:1-62Plant phenotyping relevance match · UnverifiedOpenAlex · checked 14 Sept 2026
Assessment of chlorophyll content is important to understand plant nitrogen status in precision agriculture. Traditional destructive methods for chlorophyll quantification are time-consuming, labor-intensive, and unsuitable for high-throughput phenotyping applications. The SPAD meter (Soil Plant Analysis Development) provides a rapid and non-destructive alternative by measuring leaf greenness as a proxy for chlorophyll content. Recent technological advances in imaging sensors and computational methods have enabled the development of low-cost approaches for predicting SPAD values. In this study, we propose an ensemble deep learning model-based Android application ( SPAD Predictor ) that was developed for predicting the SPAD value from RGB contact imaging. A total of 34 features, including color space features, RGB-derived features, and vegetation indices, were used to develop the model. The model consists of a lightweight Multi-Layer Perceptron (MLP) and Random Forest (RF) Regressor layer with stacking ensemble architecture. A linear regression was used as a meta-model to ensemble the MLP and RF layers. A permutation-based feature importance analysis showed that the a* channel, ExGR, RG, NRI and VARI indices played the most important roles in predicting SPAD value. The proposed ensemble deep learning model yielded R 2 t r a i n i n g of 0.987 and R 2 t e s t i n g of 0.89, RMSE of 3.25. The developed application was successfully deployed and was able to perform image submission, backend communication, prediction generation, result display, and history management. Field-level validation of the developed application yielded R 2 of 0.848, RMSE of 3.068, and MAE of 2.544. These findings indicate that the developed system has practical potential as a low-cost, field-applicable tool for estimating paddy leaf SPAD.
Why it matches plant phenotyping methodsRGB画像からイネ葉のSPAD値を推定するアプリと深層学習モデルを開発し、フィールド検証も実施しており、植物表現型取得手法が研究の中心である。
abstractwe propose an ensemble deep learning model-based Android application ( SPAD Predictor ) that was developed for predicting the SPAD value from RGB contact imaging.
ABSTRACT The common bean is vital for food security, but its productivity is often limited by competition with weeds, requiring the use of herbicides. The response of genotypes to herbicides such as fomesafen and imazamox is variable, and the traditional evaluation of phytotoxicity through visual methods is subjective. Therefore, the present study aimed to: (i) propose a methodology based on image analysis for phenotyping herbicide‐induced phytotoxicity in common bean genotypes, aiming to reduce the subjectivity of traditional visual assessments; (ii) characterize common bean genotypes under the effects of different herbicides and their doses in progenies and parental lines, based on morphophysiological traits and indices derived from visible RGB (red, green, and blue) digital images. The experiment was conducted in a completely randomized design under a 3 × 3 × 6 factorial scheme (herbicide × dose × genotype) with three replications, evaluating fomesafen and imazamox at doses of 0%, 100%, and 200% of the recommended rates. Data were collected on visual phytotoxicity, plant height, stem diameter, number of leaves, and image indices (Green Index, Excess Green Index, Excess Red Index, and Color Index of Vegetation Extraction). Results indicated that the triple interaction was significant, revealing the complexity of plant responses to herbicides. Canonical discriminant analysis explained 78.66% of the total variation, with the first canonical discriminant function (29.42%) contrasting structural development and vitality with stress, the second canonical discriminant function (27.59%) reflecting overall plant vigor, and the third canonical discriminant function (21.65%) capturing stress and phytotoxicity negatively affecting growth. The analysis demonstrated that image‐based indices combined with multivariate techniques are effective for quantifying phytotoxicity and distinguishing genotypes (tolerant and sensitive to herbicide effects), overcoming the limitations of visual evaluations, and should be used as a complementary tool to traditional techniques. Therefore, the parental genotype IPR Campos Gerais and the progeny F1A were tolerant to herbicides at different doses, while the parental genotype BAF36 and the progeny F2B were sensitive. Hence, the proposed methodology is effective for identifying herbicide‐tolerant and sensitive genotypes.
Why it matches plant phenotyping methodsRGB画像解析と多変量解析による除草剤誘発 phytotoxicity の表現型評価法の提案が研究の中心であり、従来の主観的評価を改善する方法開発に該当する。
abstractthe present study aimed to: (i) propose a methodology based on image analysis for phenotyping herbicide‐induced phytotoxicity in common bean genotypes
Rice (Oryza sativa L.) is a crucial food crop, supplying a significant portion of the global population's caloric intake. With the shift from traditional breeding methods to digital approaches, image analysis is becoming essential for distinguishing between rice cultivars. However, the optimal growth stages for effectively utilizing image analysis to classify rice varieties remain uncertain. This study aimed to evaluate 102 rice cultivars through non-destructive image processing and RGB ratio analysis. Images were captured every two days throughout the growth period, and an RGB ratio formula was developed, excluding background pixels to focus on plant characteristics. Regression analysis identified critical time points for differentiation, with the red (R) channel being most effective at 55 and 75 days post-transplanting, and the green (G) channel at 60 and 80 days. Hierarchical clustering of slopes from piecewise regression categorized the 102 cultivars into three distinct clusters, representing their ecological types. These findings provide a precise and efficient method for classifying rice cultivars, offering breeders key insights into the most effective stages for variety differentiation. By optimizing image analysis techniques, this research enhances the efficiency of rice breeding programs and supports the targeted management of genetic resources for improved trait selection.
Why it matches plant phenotyping methodsイネの画像から背景を除去してRGB比を算出し、時系列画像解析と回帰・クラスタリングにより品種識別に有効な時点と特徴量を開発・評価しており、植物表現型取得・抽出手法が中心である。
abstractThis study aimed to evaluate 102 rice cultivars through non-destructive image processing and RGB ratio analysis.
Urban forest health may be monitored and supported with the implementation and maintenance of an accurate community tree inventory. The longitudinal recording of a tree’s attributes (i.e. diameter and height) may inform potential inputs and activities related to maintenance, health, and the establishment of a tree protection zone. Urban tree inventories feature barriers to implementation including resources (i.e. labour, time, and finances), competing priorities, and gaps in knowledge. The objectives of this case study were to investigate the feasibility of structure-from-motion (SfM) approaches to measure tree height and trunk diameter (1) with oblique RGB imagery, (2) with and without GCP, and (3) during leaf-off and leaf-on conditions. Structure-from-Motion datasets were obtained using Unoccupied Aerial Systems (UAS) – “drones” and related components – to collect oblique aerial imagery, which was processed with photogrammetry software Agisoft Metashape. Tree height measurements using leaf-on imagery (R2 = 0.58, RMSE = 1.34 m) were more accurate when compared to leaf-off imagery with Ground Control Points (GCP) (R2 = 0.47, RMSE = 3.41 m) and leaf-off imagery without GCPs (R2 = 0.43, RMSE = 3.49 m). Tree height measurements during the leaf-off period had no significant difference when comparing imagery with and without GCPs. Trunk diameter measurements using leaf-off imagery were not significantly different with the use of GCPs (R2 = 0.68, RMSE = 6.39 cm) compared to those without (R2 = 0.68, RMSE = 7.85 cm). This case study highlights the applicability and accuracy of Unoccupied Aerial Systems and Structure-from-Motion methods when collecting important urban tree inventory parameters, and presents an accessible, reliable, and replicable workflow for urban forestry practitioners with limited photogrammetry-related experience.
Why it matches plant phenotyping methodsUAS-SfMフォトグラメトリによる樹高・幹径という植物形態形質の推定法を開発・比較検証し、精度と再現可能なワークフローを評価しているため、方法が中心的です。
abstractThe objectives of this case study were to investigate the feasibility of structure-from-motion (SfM) approaches to measure tree height and trunk diameter
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · Crossref · checked 5 Sept 2026
Abstract Precision field management and high-throughput plant phenotyping increasingly rely on remote sensing to capture spatial and temporal variability in crop performance. Unmanned aerial vehicle (UAV) – based sensing offers unique advantages for field-scale data collection, including high spatial resolution, flexible deployment, and scalable throughput. However, the full potential of UAV platforms remains constrained by labor-intensive operations across flight execution, data transfer, and processing workflows. This study presents a systematic evaluation of an automatic UAV-based crop sensing platform through a season-long, multi-crop field experiment. Data acquisition was conducted over a maize irrigation trial and a soybean breeding experiment, resulting in 176 completed flights over 28 days during the growing season. High-frequency flights on selected days captured diurnal dynamics in key canopy traits, including maize leaf rolling under drought stress and genotype-dependent plot temperature variation in soybean. In the maize irrigation experiment, significant differences in diurnal canopy cover ratio (CCR) were observed among irrigation treatments. The predictive relationship between CCR and final grain yield strengthened throughout the day, with the coefficient of determination (R 2 ) increasing from 0.05 in the early morning (RMSE = 3.05 Mg ha − 1 ) to 0.65 at midday (RMSE = 1.87 Mg ha − 1 ), highlighting the importance of temporal optimization in UAV-based sensing. Temperature measurements from the onboard thermal infrared camera showed a strong overall linear correlation with ground truth measurements (R 2 = 0.85). In the soybean trial, the highest plot temperature was observed on the fast-wilting genotype. Additionally, regression models were developed to estimate key crop traits, including canopy height (CH) and leaf area index (LAI), demonstrating the platform’s quantitative sensing capability. Overall, this study demonstrates that automatic UAV systems enable high-temporal-resolution crop monitoring while substantially reducing operational cost. The results highlight their potential for precise crop management and scalable field phenotyping. Future work will focus on integrating automated data processing pipelines to support near-real-time analytics and decision-making.
Why it matches plant phenotyping methods自動UAVのRGB・熱画像センシング platform を圃場で系統的に評価し、温度・キャノピー被覆率・高さ・LAIなどの植物形質を定量化しているため、フェノタイピング手法が研究の中心である。
abstractThis study presents a systematic evaluation of an automatic UAV-based crop sensing platform through a season-long, multi-crop field experiment.
Plant diseases threaten global agriculture, causing 20–40% yield losses and food insecurity. Current diagnostic methods are costly and lack scalability. While deep learning advances plant disease detection, there remains a need for CNNs with simpler architectures, better generalizability, and lower computational cost. This study presents a novel CNN for multi-class classification of 38 diseases. Trained on a public dataset of over 87,000 RGB images, the architecture comprises five convolutional blocks (filters 32–512) with max pooling and dropout (0.25, 0.4), followed by a 1,500-unit dense layer and SoftMax output. Optimized with Adam (lr=0.0001) and categorical cross-entropy, the model achieved 98% training and 96% validation accuracy with approximately 28.7 million parameters significantly fewer than transfer learning architectures. These results demonstrate an effective balance between predictive performance and computational efficiency, positioning the model as a promising tool for real-world agricultural deployment.
Why it matches plant phenotyping methods葉画像から植物病害を分類するCNNの開発と性能評価が研究の中心であり、植物の病害状態を直接推定するため、植物フェノタイピング手法に該当します。
titleLeaf-by-Leaf Diagnosis: A Custom CNN with Pyramidal Feature Extraction for Plant Disease Classification
Reproduction assets foundThe paper's sole qualifying asset is the plant disease image dataset used for all its CNN training/validation measurements: the publicly available New Plant Diseases Dataset (Augmented) on Kaggle, explicitly declared in the Data availability statement. No author code, trained model checkpoints, or other paper-specific Dataset · publict to disclose.
Acknowledgment:
The authors are sincerely grateful to their institutions for their continued support and trust, which greatly contributed to
the completion of this research.
Data availability
The dataset used and analyzed during the current study, “New Plant Diseases Dataset”, is publicly available on Kaggle at:
(https://www.kaggle.com/datasets/vipoooool/new-plant-diseases-dataset?select=New+Plant+Diseases+Dataset%28Augmented%29)
References
[1] P. Arputharaj and K. Karunanithy, "A review on machine learning and deep learning techniques for plant leaf disease
detection and classification with IoT in agriculture industry," Journal of Industrial Information Integration, vol. 50, Open asset ↗Kaggle · vipoooool/new-plant-diseases-datasetpdf-raw-page:11 lines:1-49Dataset · publict to disclose.
Acknowledgment:
The authors are sincerely grateful to their institutions for their continued support and trust, which greatly contributed to
the completion of this research.
Data availability
The dataset used and analyzed during the current study, “New Plant Diseases Dataset”, is publicly available on Kaggle at:
(https://www.kaggle.com/datasets/vipoooool/new-plant-diseases-dataset?select=New+Plant+Diseases+Dataset%28Augmented%29)
References
[1] P. Arputharaj and K. Karunanithy, "A review on machine learning and deep learning techniques for plant leaf disease
detection and classification with IoT in agriculture industry," Journal of Industrial Information Integration, vol. 50, Open asset ↗Kaggle · vipoooool/new-plant-diseases-datasetpdf-raw-page:11 lines:1-49Code / dataset availability confirmedEurope PMC · Crossref · checked 5 Sept 2026
Abstract Plant diseases affecting leaves and fruits cause substantial yield and economic losses worldwide, particularly in horticultural crops cultivated under diverse agro-climatic conditions. Early and accurate disease diagnosis is essential for effective crop management; however, manual inspection is time-consuming, subjective, and often infeasible at large scale. In this work, we present the Tomato–Chilli–Papaya (TCP) Fruit and Leaf Disease Dataset, a comprehensive multi-crop image dataset designed to support deep learning-based plant disease recognition. The dataset comprises labeled RGB images of healthy and diseased leaves and fruits from three economically important crops—tomato, chilli, and papaya—captured under real-field and semi-controlled environments, reflecting significant variability in illumination, background complexity, and disease severity. To demonstrate the applicability of the dataset, several commonly used convolutional neural network (CNN) architectures, including VGG, ResNet, DenseNet, MobileNet, and EfficientNet models, were trained and evaluated on the TCP dataset using transfer learning. Experimental results show that deep CNN models can effectively learn discriminative visual features corresponding to disease-specific patterns such as leaf spots, lesions, discoloration, curling, and fruit surface abnormalities. Lightweight models such as MobileNet achieve competitive performance with reduced computational cost, while deeper architectures provide improved accuracy at the expense of higher complexity. The results highlight the importance of dataset diversity for robust model generalization across multiple crops and plant organs. The TCP dataset provides a challenging benchmark for single-crop and multi-crop disease classification and supports the development of advanced deep learning, attention-based, and explainable AI models for precision agriculture. By enabling reproducible research and realistic performance evaluation, this dataset contributes toward scalable and practical AI-driven plant disease diagnosis systems aimed at reducing yield losses and supporting sustainable agriculture.
Why it matches plant phenotyping methods植物の葉・果実の病徴を画像から評価する大規模データセットとベンチマークを中心に扱っており、植物病害状態の画像ベース表現型解析に該当する。
abstractwe present the Tomato–Chilli–Papaya (TCP) Fruit and Leaf Disease Dataset, a comprehensive multi-crop image dataset designed to support deep learning-based plant disease recognition.
Reproduction assets foundThe paper introduces the TCP (Tomato-Chilli-Papaya) fruit and leaf disease image dataset and reports CNN experiments on it. The dataset is publicly deposited on Mendeley Data, and the authors state that analysis code is available on GitHub. Both are paper-specific, public, and actionable.Dataset · publicData is available on Mendeley:1Open asset ↗pdf-page:27 lines:1-51Code · publicCode availability: Code is available on GitHub 2Open asset ↗GitHubpdf-page:27 lines:1-51Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · Europe PMC · checked 5 Sept 2026
Introduction: Plant phenotyping requires accurate and repeatable three-dimensional structural information, but practical acquisition conditions in greenhouses, seedling rooms, and indoor pot experiments often include complex backgrounds, handheld motion blur, and thin leaf structures. These factors reduce the robustness of conventional three-dimensional reconstruction methods and limit their use in low-cost and automated phenotyping. Methods: To address this problem, this paper proposes F2DMAS, an automated three-dimensional plant phenotyping workflow using consumer-grade smartphone videos. The workflow first converts multiview RGB videos into image sequences and removes motion-blurred frames through frequency-domain quality filtering. A frequency-spatial plant segmentation module, termed FSAM3, is then introduced to separate plant structures from complex backgrounds without task-specific annotated training data. The segmented image sequences are further reconstructed using 2D Gaussian Splatting, followed by TSDF-based meshing, scale recovery, and virtual measurement for extracting plant height, canopy width, leaf length, and leaf width. Results: Experiments were conducted on 15 plant species under two acquisition scenarios. The proposed workflow achieved stable plant reconstruction under non-ideal background conditions, with PSNR, SSIM, and LPIPS values of 31.09, 0.9711, and 0.0365, respectively. Compared with the baseline reconstruction workflow, F2DMAS substantially reduced the processing time for mesh extraction while improving reconstruction quality. The extracted phenotypic traits showed strong agreement with manual measurements, with R² values ranging from 0.90 to 0.99, RMSE values ranging from 0.64 to 1.21 cm, and MAPE values ranging from 4.50% to 9.73%. Discussion: These results indicate that F2DMAS can provide an end-to-end workflow from smartphone video acquisition and plant segmentation to three-dimensional reconstruction and phenotypic trait extraction. The proposed method offers a practical and deployable solution for greenhouse seedling cultivation, potted plant experiments, and low-cost three-dimensional plant phenotyping.
Why it matches plant phenotyping methodsスマートフォン動画から植物の3D構造を再構成し、複数の形態形質を抽出・検証するワークフロー自体が中心的な方法論的貢献である。
abstractThe extracted phenotypic traits showed strong agreement with manual measurements, with R² values ranging from 0.90 to 0.99
Hybrid maize performance depends strongly on the genetic purity of hybrid seeds, but female self-pollinated seeds and target hybrid seeds are difficult to distinguish by conventional visual inspection because of their highly similar phenotypes. This study developed a nondestructive and interpretable maize hybrid purity detection framework by integrating hyperspectral imaging and RGB-derived texture features. Hyperspectral images were acquired from the embryo and endosperm sides of five female parents, one common male parent, and their corresponding hybrids. Texture-based, full-band spectral, characteristic-band spectral, and texture-spectral fusion models were systematically constructed and compared. Competitive Adaptive Reweighted Sampling (CARS), Successive Projections Algorithm (SPA), and Synchronous Two-Dimensional Correlation Spectroscopy (Sync2D) were used for characteristic wavelength selection. The results showed that the embryo side provided more stable and discriminative spectral information than the endosperm side. Texture-only models showed limited ability to distinguish hybrids from female self-pollinated seeds, whereas embryo-side texture-spectral fusion models combined with CARS or SPA and Support Vector Machine (SVM) or Partial Least Squares Discriminant Analysis (PLS-DA) achieved average test accuracies of 0.99-1.00, meeting the national maize hybrid seed purity requirement of 97%. In the optimal low-dimensional models, the retained high-dimensional spectral variables were compressed to 28-69 key features, corresponding to a dimensionality reduction ratio of approximately 88%-95%. SHAP analysis identified mean saturation and seed size as important texture features; among spectral intervals, the 450-462 nm region appeared among the top-ranked embryo-side SHAP features in all five maize lines and showed the highest embryo-side mean absolute SHAP magnitude (0.0107 ± 0.0028). Overall, the proposed framework provides a high-throughput, low-dimensional, and interpretable solution for maize hybrid seed purity detection.
Why it matches plant phenotyping methodsハイパースペクトル画像、RGBテクスチャ、特徴選択、機械学習を統合し、トウモロコシ種子のハイブリッド純度を非破壊推定する手法の開発・比較が研究の中心である。
abstractThis study developed a nondestructive and interpretable maize hybrid purity detection framework by integrating hyperspectral imaging and RGB-derived texture features.
Accurate estimation of crop yield and biomass using UAV-based remote sensing is influenced by flight altitude, plant density, canopy structure, vegetation index (VI) selection, and crop growth stage. This study evaluated how the joint optimization of these factors influences biomass and yield prediction in common buckwheat (Fagopyrum esculentum Moench var. 'Yangjeol'). Field experiments were conducted using broadcast seeding and drill seeding at four row spacings (12.5, 20, 30, and 40 cm) following the complete Latin square design. UAV RGB imagery was acquired at the third-leaf and full-flowering stages from 30, 50, and 60 m altitudes, and vegetation indices Excess Green Index (ExG), Green Leaf Index (GLI), and Normalized Green-Red Difference Index (NGRDI)were extracted. ANOVA with Tukey's HSD revealed significant variations in biomass and yield traits among sowing treatments (p < 0.005). The highest fresh weight was recorded under drill seeding at 12.5 cm spacing, while seed weight was consistently higher under all drill seeding treatments compared with broadcast seeding. The number of seeds per plant peaked under 40 cm spacing, indicating a trade-off between planting density and reproductive output. Strong and significant correlations between vegetation indices and ground-measured traits were observed (r = 0.82-0.98), but these relationships were highly dependent on growth stage, sowing configuration, and UAV altitude. The third-leaf stage under broadcast seeding and full flowering stage under drill seeding at 20 cm spacing showed the strongest and most consistent VI-trait associations. Among UAV altitudes, 50 m provided the most stable predictive performance across traits. ExG and GLI exhibited more consistent relationships with biomass and yield parameters than NGRDI. These findings demonstrate that no single UAV altitude, vegetation index, or growth stage is universally optimal. Instead, coordinated optimization of UAV operational parameters and sowing configurationsubstantially improves the reliability of UAV-based yield and biomass estimation in buckwheat.
Why it matches plant phenotyping methodsUAV RGB画像と植生指数を用いたバイオマス・収量推定の条件最適化と精度評価が研究の中心であり、植物形質の取得手法を実質的に検証している。
abstractAccurate estimation of crop yield and biomass using UAV-based remote sensing is influenced by flight altitude, plant density, canopy structure, vegetation index (VI) selection, and crop growth stage.
The segmentation model achieved Mean IoU values of 0.7723 for Water Stress 2025, 0.9164 for Common Rust 2025, and 0.9531 on the benchmark dataset. The classifier achieved 99.54% accuracy for the five-class task; however, the improvement over the strongest baselines and the RGB + multispectral configuration was limited. Therefore, the classification component is not presented as a substantially superior classification-only model. Instead, it is interpreted as an exploratory multimodal analysis that quantifies the contribution and limitation of RGB, multispectral, Wavelet, and GLCM branches under the adopted UAV dataset protocol. For classification-only deployment, simpler alternatives such as DenseNet201 or the RGB + multispectral configuration may be more practical because they provide comparable accuracy with lower architectural or preprocessing complexity. Ablation, modality-controlled, and 21-run stability experiments showed reproducible segmentation results and clarified the behavior of the classification branches. RGB and multispectral branches mainly provided the peak classification accuracy, whereas Wavelet and GLCM branches mainly affected offline convergence rather than final accuracy. RGB, NDVI, and NDRE visualizations were also added for qualitative support. Since direct physiological ground measurements were not available for all samples, the masks are interpreted as adaptive index-guided labels rather than direct physiological ground truth. Overall, the main evidence of practical benefit is associated with UAV-based dataset construction, adaptive index-guided segmentation, and field-scale stress/disease mapping, while the classification experiments should be interpreted as modality-contribution and convergence analyses rather than proof of a practically superior complex classifier.
Why it matches plant phenotyping methodsUAVマルチスペクトル画像によるトウモロコシの水ストレス・病害状態のセグメンテーション、データセット構築、再現性評価が研究の中心であり、植物表現型の取得・抽出手法として実質的です。
abstractOverall, the main evidence of practical benefit is associated with UAV-based dataset construction, adaptive index-guided segmentation, and field-scale stress/disease mapping
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · bioRxiv · checked 5 Sept 2026
Abstract In plant breeding and genetics, recent advances in high-throughput phenotyping are beginning to meet the growing demand for large-scale, high-quality phenotypic data that emerged after the development of next-generation sequencing technologies. Recent developments in phenomics have been incorporated into almond breeding programs, facilitating the large-scale acquisition of quantitative phenotypes and the dissection of the genetic architecture underlying morphological and quality-related traits. The implementation of a high-throughput phenotyping platform integrating RGB and hyperspectral imaging with genotyping using the 60K almond SNP array enabled the large-scale characterization of almond populations and the identification of 567 robust marker–trait associations across 66 traits. These analyses revealed two major genomic hotspots on chromosomes 2 and 5 associated with morphological and quality-related traits. These regions harbored biologically relevant candidate genes, including genes associated with OVATE family proteins, brassinosteroid signaling, protein ubiquitination, and acyl-CoA metabolism, as well as other regulators of organ growth, cell proliferation, hormone signaling, and seed development. Furthermore, a novel candidate gene encoding a COMT-like O-methyltransferase involved in lignin biosynthesis was identified and proposed to contribute to shell hardness, a major genetically controlled trait in almond. Together, these findings demonstrate the potential of integrating high-throughput phenomics and genomics to dissect complex traits, identify candidate genes, and accelerate genomics-informed breeding in almond.
Why it matches plant phenotyping methodsRGB・ハイパースペクトル画像を統合した高スループット表現型解析プラットフォームの実装と大規模形質取得が研究の主要部分であり、単なる形質のルーチン測定ではない。
abstractThe implementation of a high-throughput phenotyping platform integrating RGB and hyperspectral imaging with genotyping using the 60K almond SNP array enabled the large-scale characterization of almond populations
Plants can suffer a number of diseases that impact agricultural productivity and food security, particularly in developing farming communities. Although deep learning is capable of classification of diseases with outstanding results, its use is limited due to the difficulty of obtaining large labeled databases and the high requirement of computation. To address these challenges, this study proposes a new method for plant disease classification based on traditional image processing and machine learning algorithms with lightweight and low computation requirements. This one uses several handcrafted descriptors such as color histograms, Haralick texture features and Hu moments to retrieve the information relevant to a disease from the segmented leaf images. Performance of top five classifiers, namely Random Forest, Support Vector Machine, K-Nearest Neighbors, Logistic Regression and Naïve Bayes classifiers are evaluated from the dataset of healthy plant leaves and diseased plant leaves images on 10-fold cross validation. Based on the results of the research work, the best classification model was the Random Forest Classifier model with the accuracy value is 98.12%, 0.98 precision, 0.98 recall, and 0.98 F1 value. The proposed solution was also found to be uncomputation complex and low memory consuming and can be made realtime inference. Therefore, this solution can be implemented onto agricultural systems, considering edge computing and IoT. The results of the research also demonstrated that feature-based machine learning approaches afford interpretable and reliable plant disease detection at a low computation cost, further contributing to sustainable, and precision agriculture.
Why it matches plant phenotyping methods葉画像から病徴を抽出・分類する画像処理および機械学習手法が研究の中心であり、植物の病害状態を直接推定するため、植物フェノタイピング手法として含める。
abstractthis study proposes a new method for plant disease classification based on traditional image processing and machine learning algorithms
Mango (Mangifera indica L.) is among the most important commercial fruits grown throughout the world in the tropical and subtropical areas. Even though mangoes are economically important, their cultivation is continuously threatened by a wide variety of leaf and fungi diseases, resulting in crop losses of up to 15-30% annually. [1]. Conventional disease identification depends heavily on expert visual inspection—a process that is inherently slow, subjective, and largely impractical for smallholder farmers operating in remote areas with limited access to agronomic specialists. This paper provides an end-to-end deep learning-based approach towards automatic mango leaf disease detection along with an Android application for real-time deployment in the field. This work is built upon MangoLeafBD [1], an openly accessible dataset that consists of a total of 4,000 images of RGB color space for seven different disease categories and one healthy class, with each category having 500 samples collected from four separate orchards in Bangladesh. Three types of transfer learning models including VGG16 [2], MobileNetV2 [3], and DenseNet121 [4] were considered after applying two-phase fine-tuning based on pre-trained ImageNet weights. For each of the three types of neural network models tested, a series of preprocessing steps consisting of bilinear resizing to size 224 x 224, channel-wise normalization, and image augmentation (rotation, zoom, brightness adjustment, horizontal flip, and shear) was used to increase model accuracy for diverse real-world images. The training model was deployed using TensorFlow Lite (TFLite), which allowed it to be run in offline mode on mid-end Android phones, without needing internet connectivity. The app lets farmers upload leaf images or take images and get a diagnosis of the leaf diseases, along with possible treatments for them.This work contributes a replicable pipeline linking state-of-the-art deep learning research with practical precision agriculture, particularly for rural communities that currently lack access to timely agronomic advisory services.
Why it matches plant phenotyping methodsマンゴー葉画像から病害状態を推定する深層学習パイプラインとモバイル実装が研究の中心であり、植物病害フェノタイピング手法に該当する。
abstractThis paper provides an end-to-end deep learning-based approach towards automatic mango leaf disease detection along with an Android application for real-time deployment in the field.
Rice seedling salt-tolerance evaluation commonly relies on visual scoring or destructive assays, which are subjective, labor-intensive, and difficult to standardize for population-level analysis. This study developed a new deep-learning-assisted hydroponic RGB phenotyping framework for standardized salt-stress evaluation and genetic mapping in rice seedlings. The framework integrates controlled hydroponic cultivation, RGB imaging, RicePhenoSeg-assisted annotation and trait extraction, ELMERF-based semantic segmentation, and image-derived quantification of salt-induced shoot injury. Using this framework, we constructed the Rice Seedling-Salt RGB Dataset (RSSD), which contains green shoot tissues, yellow shoot tissues, roots, and background from hydroponically grown rice seedlings. Based on RSSD, ELMERF achieved a mean Intersection over Union of 51.4% and a mean Accuracy of 89.5%, outperforming nine representative segmentation models. We further defined shoot yellowing rate (SYR) as an image-derived quantitative trait describing visible salt-induced shoot injury. The framework was applied to 261 re-sequenced rice accessions for population-level phenotyping and genome-wide association analysis. Compared with standard evaluation score and seedling death rate, SYR showed a more continuous phenotypic distribution and detected 36 significant SNPs, including a major signal near the Saltol/OsHKT1; 5 region. Notably, 34 SYR-associated SNPs were not detected by conventional visual scores. Overall, this study provides a targeted hydroponic RGB phenotyping framework for standardized rice seedling salt-stress evaluation and genetic analysis.
Why it matches plant phenotyping methods深層学習によるRGB画像セグメンテーション、形質抽出、データセット構築、性能比較を中核とし、画像由来の塩ストレス傷害形質を定量化する植物フェノタイピング手法である。
abstractThis study developed a new deep-learning-assisted hydroponic RGB phenotyping framework for standardized salt-stress evaluation and genetic mapping in rice seedlings.
Reproduction assets foundThe paper's Data Availability statement explicitly deposits datasets, source code, and supporting data in a public GitHub repository (ELMERF), which covers the RSSD RGB image dataset, segmentation code, and phenotyping/GWAS analysis assets. RiceVarMap is a cited external SNP database, not a paper-specific asset.Code · publicThe datasets, source code, and other supporting data are openly available on the ELMERF repository (https://github.com/PhenoCodexh/ELMERF).Open asset ↗PhenoCodexh/ELMERFhtml-lines:446-478Plant phenotyping relevance match · UnverifiedOpenAlex · arXiv · checked 5 Sept 2026
High-throughput plant phenotyping generates valuable data that often remains trapped in unstructured text and isolated RGB images. To bridge this semantic gap, we propose a framework for constructing a multimodal granular Knowledge Graph (KG) to monitor genotype-phenotype interactions across time and experiments. In this work, we focus on wheat Triticum aestivum as a representative target crop to validate our methodology across complex canopy environments. Our pipeline first distills noisy field notes to extract entities and relations, dynamically constructing the KG by converting unique instances into hierarchical class entities via RDF-typing. These graph nodes are then aligned with standardized ontologies (PO, RO, WTO) using PlantDeBERTa. To visually ground the constructed graph, a Vision-Language Model paired with a wheat-segmentation ViT generates attention-based softmaps, linking specific KG entities directly to image pixels. We introduce a central observation node Plant_Obs_Id to connect these multimodal subgraphs temporally. Evaluated on 500 curated WisWheat samples using Pointing Game accuracy, Visual Word Sense Disambiguation (VWSD), and rank-based metrics, our neuro-symbolic approach successfully maps complex field observations to a structured graph. This enables automated field note auditing, temporal stress monitoring, and precise spatial trait localization for wheat breeders.
Why it matches plant phenotyping methods植物のマルチモーダル表現型データを知識グラフと画像に統合し、画像画素への形質局在化を行う中核的な計算フレームワークを提案・評価しているため。
abstractwe propose a framework for constructing a multimodal granular Knowledge Graph (KG) to monitor genotype-phenotype interactions across time and experiments
Plant diseases remain a major challenge to global food production, and timely, accurate, and scalable detection of plant stress is critical to reducing these losses. Recent advances in digital imaging and artificial intelligence offer unprecedented opportunities for precision crop disease detection and management. Yet, existing plant disease datasets remain often fragmented across crop and disease systems, and are largely dominated by controlled-environment imagery. The lack of standardized, interoperable, and representative datasets limits reproducibility, transferability, and scalability of AI systems, thereby constraining their deployment in operational agricultural applications. Here we present LeafMD, an integrated multimodal plant disease dataset and benchmark resource that includes LeafNet 2.0, a large-scale multimodal digital image dataset comprising 255,855 image–text pairs across 37 crop species, 197 crop–disease classes, and 9 geographic regions spanning tropical, subtropical, and temperate agricultural systems. Unlike conventional datasets, LeafNet 2.0 integrates biologically grounded symptom descriptions with image-level annotations of early and late disease stages, enabling symptom-aware analysis of disease progression under realistic field conditions. We further introduce LeafBench 2.0 as part of LeafMD, a visual-question answering benchmark covering nine fine-grained plant pathology tasks, including pathogen classification, lesion characterization, symptom interpretation, and disease severity assessment. Evaluation across 16 vision–language models revealed substantial performance gaps between coarse disease recognition and fine-grained pathological reasoning, while agriculture-adapted models consistently outperformed several larger general-domain architectures on symptom-oriented tasks. Together, LeafNet 2.0 and LeafBench 2.0 establish LeafMD as a multimodal resource for developing disease-aware agricultural foundation models and studying fine-grained pathological reasoning in real-world environments.
Why it matches plant phenotyping methods植物病害の画像・症状記述データセットとベンチマークを構築し、病徴解釈・病変特徴・病害重症度評価を対象にモデル性能を評価しており、植物状態の取得・評価手法が中心である。
abstractHere we present LeafMD, an integrated multimodal plant disease dataset and benchmark resource
Accurate monitoring of nitrogen nutrition is critical for optimizing cotton production. Traditional machine learning-based inversion models have limited effectiveness for precision monitoring. Multisource fusion models for small samples were developed in this study to achieve enhanced accuracy through fitting and data complementarity. Cotton plants subjected to different nitrogen treatments were investigated. A two-year pot experiment was conducted to collect main-stem leaf images to construct an image pretraining dataset for model transfer. In a field experiment conducted over one year, main-stem leaf data were collected using hyperspectral, chlorophyll fluorescence, and digital camera sources, thereby providing a multisource dataset for training monitoring models. Two architectures—a neural network (NN) and an interpretable deep forest (DF), which are suitable for small-sample spectral, fluorescence, image color, and texture-sequence features—were constructed to improve the accuracy of nitrogen content inversion. Additionally, a two-dimensional sliding-window processing method was introduced into the DF multigranularity scanning module, and a transfer-learning-based two-dimensional convolutional NN was employed to directly model small-sample two-dimensional images. Building upon the outcome, multilayer fusion models were constructed, with corresponding fusion strategies designed for homogeneous sequence inputs and heterogeneous image–sequence inputs. The results showed that NN and DF can effectively handle limited sample sizes and outperform traditional machine learning models. Among the fusion models, the optimal secondary decision-level fusion model achieved an R² of 0.926 on the independent test set, indicating good performance under small-sample conditions. This study provides a methodological reference for the precise monitoring of crop phenotypic parameters under small-sample conditions.
Why it matches plant phenotyping methods綿花葉の窒素含量という植物形質を、画像・ハイパースペクトル・蛍光データの融合と深層学習で推定する手法を開発・評価しており、形質取得・推定法が研究の中心である。
abstractMultisource fusion models for small samples were developed in this study to achieve enhanced accuracy through fitting and data complementarity.
Agriculture remains the backbone of global food security, yet crop diseases, nutrient deficiencies, water stress, and pest infestations cause annual yield losses estimated at 20–40% worldwide. Conventional field scouting methods are labour-intensive, time-consuming, and fail to capture the spatial heterogeneity of large farms. This paper presents a Drone-Based Crop Health Analysis and Precision Agriculture System (DBCHAPS) that integrates multi-spectral and RGB imaging drones, deep learning-based crop disease detection, vegetation index analysis, variable-rate prescription mapping, and autonomous precision spraying. A DJI Matrice 300 RTK drone equipped with a MicaSense RedEdge-MX multi-spectral camera captures high-resolution aerial imagery across five spectral bands (Blue, Green, Red, Red-Edge, Near Infrared). The captured data is processed through a custom-trained YOLOv8-based convolutional neural network (CNN) pipeline to detect 18 distinct crop diseases and stress conditions across rice, wheat, and cotton crops. Concurrently, vegetation indices (NDVI, NDRE, GNDVI, SAVI) are computed to generate prescription maps for site-specific fertilizer and pesticide application. Experimental evaluation on a 120 acre farm in Thanjavur, Tamil Nadu over two crop seasons demonstrates a disease detection accuracy of 96.3%, early stress detection 8–12 days before visible symptoms, and a 31% reduction in agrochemical usage through variable-rate application. The system achieves an end-to-end field analysis time of under 45 minutes for 100 acres. Keywords — UAV, Precision Agriculture, Crop Disease Detection, Multi-Spectral Imaging, NDVI, YOLOv8, Deep Learning, Variable-Rate Application, Remote Sensing, Smart Farming.
Why it matches plant phenotyping methodsドローンのマルチスペクトル/RGB画像とYOLOv8を用いて作物の病害・ストレス状態を推定するシステムを開発し、精度と運用性能を評価しており、植物フェノタイピング手法が中心である。
abstractThis paper presents a Drone-Based Crop Health Analysis and Precision Agriculture System (DBCHAPS) that integrates multi-spectral and RGB imaging drones, deep learning-based crop disease detection, vegetation index analysis, variable-rate prescription mapping, and autonomous precision spraying.
Integrating multi-omics data, including phenomic, genomic, and environmental inputs, offers a powerful approach for enhancing maize performance and predicting grain yield. In this study, crop health was quantified using temporal NGRDI (Normalized Green Red Difference Index) trajectories collected from 16 unoccupied (unmanned) aerial vehicle or system (UAV or UAS, drones and sensors) flights (from 19 to 117 d after planting) in maize trials conducted in Texas. Crop health indices (CHIs) were calculated through area under the curve (CHIAUC) and functional principal component analysis (CHIFPCA), capturing dynamic plant health responses throughout the growing season. Heritability estimates of NGRDI fluctuated between 0.3 and 0.7, averaging 0.51 ± 0.02, reflecting consistent genetic contributions to growth dynamics. CHIs derived from a favorable (irrigated) trial in Texas effectively separated high- and low-yielding hybrids across 41 environment-tester combinations, achieving significant differentiation in 27 (CHIAUC ) and 28 (CHIFPCA1) environments. In comparison, only 21 environments were differentiated when using grain yield alone. Genomic mapping of temporal NGRDI revealed key quantitative trait loci (QTLs) linked to maize growth, containing candidate genes including br2, phyC1, wus1, mads69, cct1, rap2, miR172, and gl15, associated with canopy development, flowering regulation, and drought adaptation. Integrating multi-omics data into phenomic- and environment-informed genomic prediction models improved yield prediction accuracy by approximately 18.5%, particularly for untested genotypes in both tested and untested environments. These findings demonstrate that multi-omics integration provides a scalable framework for enhancing maize performance and advancing grain yield prediction across diverse agricultural systems.
Why it matches plant phenotyping methodsUAV画像から時系列NGRDIを抽出し、作物健康・生育動態の表現型指標を構築して検証・予測に利用しており、フェノタイピング手法の適用が研究の中心的要素です。
abstractCrop health indices (CHIs) were calculated through area under the curve (CHIAUC) and functional principal component analysis (CHIFPCA), capturing dynamic plant health responses throughout the growing season.
Accurate yield estimation and crop load monitoring are essential for precision orchard management, supporting targeted fertilization, pruning, thinning, harvest planning, and marketing decisions. However, reliable in-situ monitoring remains challenging because commercial orchards are characterized by severe canopy occlusion, fruit overlap, heterogeneous tree architecture, variable illumination, and complex backgrounds. This review synthesizes advances in multi-modal sensing and deep learning for orchard yield estimation, breaking down the paradigm into intermediate fruit-counting or crop-load monitoring steps and supplementary spectral quality-assessment dimensions. First, yield-related indicators are summarized, including direct phenotypic traits such as fruit number, size, volume, and spatial distribution, as well as indirect structural and physiological proxies such as canopy volume, vegetation indices, flowering intensity, and spectral maturity attributes. Second, representative sensing devices and carrying platforms are reviewed, including red-green-blue (RGB) cameras, red-green-blue-depth (RGB-D) sensors, light detection and ranging (LiDAR), hyperspectral and multispectral systems, unmanned ground vehicles (UGVs), and unmanned aerial vehicles (UAVs). Third, the evolution of estimation methods is discussed, from traditional image processing and machine learning to object detection, instance segmentation, multi-object tracking, point-cloud analysis, remote-sensing regression, and multi-modal fusion. The review shows that no single sensor or algorithm can satisfy all orchard monitoring requirements. Ground-based vision and depth sensing are more suitable for fine-scale fruit counting and sizing, whereas UAV and spectral sensing provide advantages for regional yield mapping and quality-enhanced assessment. Future research should emphasize occlusion-aware perception, robust cross-environment generalization, lightweight edge deployment, standardized benchmarks, and integrated quantity-quality monitoring frameworks for actionable crop load management.
Why it matches plant phenotyping methods果実数・サイズ・体積・空間分布などの植物形質を対象に、センシング機器と画像解析・深層学習による収量推定法を体系的にレビューしており、フェノタイピング手法が中心である。
abstractThis review synthesizes advances in multi-modal sensing and deep learning for orchard yield estimation, breaking down the paradigm into intermediate fruit-counting or crop-load monitoring steps and supplementary spectral quality-assessment dimensions.
The maturity level of green vegetables is an important factor affecting product quality, market value, and shelf life. Maturity identification is generally performed visually based on leaf color changes, making the assessment subjective and potentially inconsistent. This study aims to develop a classification model for green vegetable maturity levels using a combination of color feature extraction and a Convolutional Neural Network (CNN) to provide a more objective and accurate system. The research began with image acquisition of green vegetables categorized into three maturity levels: immature, mature, and overripe. Preprocessing included image resizing, normalization, and segmentation. Color feature extraction was performed using RGB and HSV color spaces to represent maturity conditions. The dataset was divided into training and testing sets with a 90:10 ratio and processed using a CNN architecture. Model performance was evaluated using accuracy, precision, recall, and F1-score. Results showed that the proposed model achieved 95.2% accuracy, 94.8% precision, 95.6% recall, and 95.1% F1-score. These findings indicate that combining color features and CNN effectively supports automated vegetable sorting and quality control systems.
Why it matches plant phenotyping methods緑色野菜の成熟度という植物器官の状態を、画像取得・色特徴抽出・CNNで自動推定する手法の開発が研究の中心である。
abstractThis study aims to develop a classification model for green vegetable maturity levels using a combination of color feature extraction and a Convolutional Neural Network (CNN) to provide a more objective and accurate system.
Drought poses a global threat to food security and demands intensified efforts from breeding programs. Yet the lack of efficient methods for selecting this trait increases the cost and time required to develop new cultivars. The goal of this work was to assess the feasibility of using spectral data from RGB or multispectral sensors for drought-tolerance classification across various machine-learning models under the most practical cross-validation scenarios typical in breeding programs. The genotypes were assessed during trials conducted under either optimal (irrigated) or drought-stress conditions across two years, and evaluated using up to 10 field traits to determine their drought-tolerance classification based on membership function values related to drought. RGB and multispectral vegetation indices collected during several flights throughout the crop cycle were used to train machine learning models. We found that drought trials offer the best training data. Specificity was the metric most affected by sensor type and the nature of the training data. The multispectral sensor outperformed the RGB sensor on most evaluation metrics in both years. AdaBoost and linear discriminant analysis models demonstrated the strongest consistency across all prediction scenarios. Together, they achieved an overall accuracy, specificity, and F1-Score of 0.71, 0.56, and 0.77, respectively. The most influential vegetation indices for model performance consistently included the NIR band. Spectral information, such as vegetation indices, is a useful tool for plant researchers to complement drought tolerance evaluations in the field. This data-driven approach facilitates automation, paving the way to speed genetic gains by including early assessments of drought tolerance in breeding pipeline, and improves resource utilization efficiency.
Why it matches plant phenotyping methodsUASのRGB・マルチスペクトルデータと機械学習を用いて干ばつ耐性を分類し、センサー比較や交差検証を行うことが研究の中心であるため、植物フェノタイピング手法として適格です。
abstractThe goal of this work was to assess the feasibility of using spectral data from RGB or multispectral sensors for drought-tolerance classification across various machine-learning models under the most practical cross-validation scenarios typical in breeding programs.
Low-cost RGB imaging is accessible for phenotyping, but color varies with devices and illumination. We tested whether RGB-derived indices from a standardized smartphone setup can proxy cotton (Gossypium hirsutum L.) leaf traits at the early seedling stage. Leaves (n=80) from three growth-chamber experiments were imaged in a closed light-tent with an in-frame gray/white/black card, then corrected in Adobe Photoshop. Mean leaf RGB values (manual ROIs) were used to compute 15 RGB/CIELAB indices, which were screened against SPAD, specific leaf area (SLA), vein density, water content (WC), stomatal density, and stomatal size using Pearson r and second-order regression (adj. R², NRMSE). The strongest relationships were for SLA (h_ab; adj. R²=0.666), vein density (TGI; adj. R²=0.610), and SPAD (G; adj. R²=0.558). WC was moderately associated with c_ab (adj. R²=0.344), while stomatal traits were weakly explained, consistent with scale limits of top-down mean-color metrics. Standardized consumer RGB imaging can therefore support rapid first-pass screening of pigment- and structure-related leaf traits.
Why it matches plant phenotyping methods標準化スマートフォンRGB撮像と色補正・指数計算を用いて葉形質を推定し、SPAD、SLA、葉脈密度などとの関係を定量評価しているため、画像フェノタイピング手法の検証が中心である。
abstractWe tested whether RGB-derived indices from a standardized smartphone setup can proxy cotton (Gossypium hirsutum L.) leaf traits at the early seedling stage.
Accurate and rapid detection of maize seedling growth is critical in early breeding decisionmaking, smart management, and yield improvement. Traditional leaf age detection still relies heavily on labor-intensive and low-efficiency manual field surveys, underscoring the urgent need for high-throughput phenotyping. Integrating multisource sensor data from unmanned aerial vehicle (UAV) with measured information such as crop height can further enhance the estimation accuracy of crop phenotypic parameters. Accurate field plot segmentation is critical for field-scale phenotypic analysis. However, current approaches remain largely dependent on slow, manual segmentation. Automating this step would greatly reduce the workload of agronomists. This study used UAV RGB and multispectral imagery collected over maize inbred line population plots before the canopy closure stage to perform automatic plot segmentation on field orthophotos and combined measured plant height with relative flight dates to achieve high-throughput detection of leaf age during the maize seedling stage. First, this study proposed a maize plot automatic segmentation method based on orthophotos. Then, it extracted texture features, RGB, and multispectral vegetation indices of each plot. Combined with relative flight date and plant height, four datasets were constructed. Support vector regression (SVR), random forest regression (RFR), and automatic machine learning (AutoML) regression algorithms were used to build the leaf age detection model. The results showed that the orthomosaic from March 23 achieved the best plot-segmentation performance, with minimum intersection over union (IoU), mean IoU, and IoU standard deviation of 14.67%, 96.47%, and 8.65%, respectively. Incorporating relative flight dates and plant-height measurements improved model performance, and the AutoML demonstrated the greatest robustness, achieving a validation R2 of up to 0.862 and an RMSE as low as 0.715. This study proposed a leaf age estimation method that offers practical technical support for field-based maize seedling assessment and reduces manual labor demands.
Why it matches plant phenotyping methodsUAV RGB・マルチスペクトル画像、圃場区画 segmentation、特徴抽出、回帰モデルを統合し、トウモロコシの葉齢という植物形質を推定する方法を開発・評価しており、フェノタイピング手法が研究の中心である。
abstractThis study used UAV RGB and multispectral imagery collected over maize inbred line population plots before the canopy closure stage to perform automatic plot segmentation on field orthophotos and combined measured plant height with relative flight dates to achieve high-throughput detection of leaf age during the maize seedling stage.
Objective This study presents an integrated, multi-scale approach for the non-destructive estimation of phenological stages and carotenoid content in carrots by combining spectral techniques, colorimetry, and artificial intelligence. Methods Six commercial varieties, including orange, yellow, white, and purple genotypes, were evaluated under field and laboratory conditions using multispectral drone imagery, high-resolution spectroradiometric signatures, red green blue (RGB) images, and CIELAB color measurements. A hierarchical modeling framework was developed across two phases: (i) spectral modeling using uncrewed aerial vehicle (UAV)-based multispectral indices, textural and geometric metrics, and laboratory-generated hyperspectral signatures; and (ii) a colorimetric index from RGB images. Results Using UAV-based multispectral field data, phenological prediction indices achieved high classification performance (F1-scores > 0.90) when modeled with a Random Forest classifier, supported by distinct spectral signatures associated with canopy development and senescence. In parallel, carotenoid content estimation using a Random Forest regression model demonstrated strong predictive accuracy ( R 2 = 0.897; RMSE = 0.584), with the Plant Senescence Reflectance Index (PSRI) and Carotenoid Reflectance Index (CRI) identified as the most influential predictors. A complementary laboratory-based Random Forest regression model using high-resolution spectral signatures achieved near-perfect predictive performance ( R 2 = 0.987). SHapley Additive exPlanations (SHAP) analysis identified physiologically relevant wavelengths in the green (540-550 nm) and red-edge (∼700 nm) regions as the primary drivers of carotenoid concentration. Likewise, a novel colorimetric index (ICarot), derived from CIELAB parameters, enabled accurate image-based carotenoid estimation ( R 2 = 0.85). Conclusion This study introduces an innovative multi-sensor framework for precision agriculture and automated postharvest quality control, enabling rapid, objective, and scalable phenotyping in carrot production systems. Through the integration of spectral, colorimetric, and AI-based approaches, the proposed methodology effectively captures both internal nutritional attributes and external quality traits within a unified, non-destructive assessment pipeline.
Why it matches plant phenotyping methods複数センサー画像・分光計測とAIを統合し、ニンジンの生育段階およびカロテノイド含量を非破壊推定する手法を開発・評価しており、表現型取得が研究の中心である。
abstractThis study presents an integrated, multi-scale approach for the non-destructive estimation of phenological stages and carotenoid content in carrots by combining spectral techniques, colorimetry, and artificial intelligence.
Reproduction assets foundThe paper's Data Availability section explicitly deposits the study's data (and project materials) on GitHub and Zenodo, both with authors' public URLs matching allowed_urls. These qualify as paper-specific public assets for the carrot phenotyping measurements and analysis.Dataset · publicThe data is available at GitHub and Zenodo:
- https://github.com/agrocompuepidemlab/Carrot-value-chain-proyect/tree/mainOpen asset ↗github.com/agrocompuepidemlab/Carrot-value-chain-proyectlines:184-307Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
Automated rice seed vigor classification provides a non-invasive and scalable solution for improving agricultural decision-making. This study proposed an image-based framework to compare traditional machine learning and deep learning approaches for classifying individual rice seed vigor using standard RGB images. Machine learning models were developed using hand-crafted morphological and color features, while convolutional neural networks were employed to automatically extract visual patterns related to seed quality. Both single-time-point and multi-time-point image analysis strategies were investigated. Models trained on images captured at individual growth stages were compared with a multi-time-point ensemble approach that integrated visual information across multiple developmental stages. The ensemble approach achieved superior performance, highlighting the importance of incorporating temporal growth dynamics into vigor classification. Notably, traditional machine learning models performed comparably to deep learning models when informative features were carefully engineered. To improve transparency and reliability, interpretability techniques were applied to better understand model decisions. Overall, the findings demonstrate the practical potential of data-driven, image-based seed vigor assessment.
Why it matches plant phenotyping methodsRGB画像と機械学習・深層学習を用いてイネ種子の活力を自動推定する枠組みを開発・比較しており、表現型取得・抽出手法が研究の中心である。
abstractThis study proposed an image-based framework to compare traditional machine learning and deep learning approaches for classifying individual rice seed vigor using standard RGB images.
Accurate crop damage detection is vital for sustaining agricultural productivity and ensuring food security, particularly within the scope of precision farming. Despite significant advancements in Deep Learning (DL), the classification of crop damage under diverse real-world conditions remains challenging due to heterogeneous damage patterns, variable crop types, and environmental inconsistencies. To overcome these challenges, a hybrid DL framework (i.e., EnCropNet) is proposed for binary crop damage classification using field-level RGB images. The proposed model integrates the global semantic representation power of DenseNet121 with a lightweight convolutional stream inspired by ShuffleNetV2, enhanced via Squeeze-and-Excitation (SE) blocks for improved channel-wise attention. The model was trained and evaluated on a balanced version of the publicly available CGIAR Crop Damage Classification (CDC) dataset. To enhance generalization, extensive augmentation techniques such as random rotations, brightness variations, and zoom transformations were applied. The proposed model outperformed leading baseline models, including LightCDC, DenseNet121, EfficientNetV2S, and ShuffleNetV2, achieving a notable test accuracy of 90.00%. Additionally, visualization tools such as GradCAM and t-SNE confirm EnCropNet ability to capture discriminative features while maintaining transparency. The results suggest that EnCropNet is an effective and scalable solution for real-time crop damage assessment, particularly in low-resource agricultural environments. Its deployment could support timely decision-making and contribute to sustainable farming practices.
Why it matches plant phenotyping methods圃場RGB画像から作物被害状態を推定する深層学習手法を開発し、公開データセットで性能比較・評価しており、植物フェノタイピング手法が中心です。
abstracta hybrid DL framework (i.e., EnCropNet) is proposed for binary crop damage classification using field-level RGB images.
Accurate and non-destructive volume estimation of agricultural products is essential for precision agriculture, yet remains challenging when transitioning from controlled laboratory conditions to complex orchard environments. Although 2D image-based volume estimation methods provide a cost-effective and scalable solution, existing studies are fragmented and lack a unified perspective on their real-world applicability. This review presents a systematic synthesis of 2D image-based volume estimation methods, explicitly framed through the laboratory-to-orchard transition. We categorized existing volume estimation approaches according to the sensing modality into monocular RGB-based approaches and depth-assisted methods, and further reviewed them based on the image processing methods. A key finding is that high-precision geometric estimation can be achieved in laboratory environments, whereas deep learning and RGB-D fusion have driven a shift from conventional geometric modeling toward data-driven and hybrid learning frameworks in orchard settings. However, 2D image-based volume estimation remains fundamentally limited by scale ambiguity, severe occlusion, and sensitivity to illumination and background variability in real orchard environment. Overall, this review provides a unified perspective for understanding volume estimation methodology across environments and offers guidance for developing robust, scalable, and field-deployable volume estimation systems for real-world agricultural applications.
Why it matches plant phenotyping methods農産物の2D画像から体積を推定する手法を体系的にレビューしており、植物器官・果実の形態形質取得方法が中心である。
abstractThis review presents a systematic synthesis of 2D image-based volume estimation methods, explicitly framed through the laboratory-to-orchard transition.
PHENET's Use Case 1 on plant health is validating sensors and imaging methods for the assessment of wheat ear diseases, with two AI-powered applications now reaching proof-of-concept stage. The first, FUSASEYD, addresses Fusarium Head Blight (FHB), a major fungal disease in winter wheat. Using RGB field images and a deep learning instance segmentation model (YOLOv11), the application detects and quantifies FHB symptoms on wheat ears, offering an automated alternative to time-consuming expert visual scoring. GEVES has developed both a PC interface and a smartphone application to visualise model predictions in the field. Validation in French registration trials is planned for the 2026 campaign. A companion article by V. Cadot et al. is currently under review in the Journal of Experimental Botany special issue on Plant Phenomics & Enviromics Across Scales. The second, COYL (Counting Orange and Yellow Larvae), tackles a practical challenge faced by breeders, and rapidly counting wheat blossom midge larvae, both Sitodiplosis mosellana and Contarinia tritici, to characterise variety susceptibility. Using smartphone RGB images and YOLOv-based object detection, the best-performing model achieved high accuracy and successfully distinguished between the two visually similar species. An online counting application has been developed, currently accessible to Walloon Agricultural Research Centre members. The labelled COYL-1 dataset is publicly available at https://doi.org/10.5281/zenodo.19402333 for community use. A companion article by Antoine Deryck et al. is under submission at Plant Phenomics Journal.
Why it matches plant phenotyping methodsRGB画像と深層学習によるコムギ穂の病徴検出・定量化を開発し、専門家評点の自動化とセンサー／画像法の検証を目的とするため、植物フェノタイピング手法が中心である。
abstractPHENET's Use Case 1 on plant health is validating sensors and imaging methods for the assessment of wheat ear diseases
PHENET's Use Case 1 on plant health is validating sensors and imaging methods for the assessment of wheat ear diseases, with two AI-powered applications now reaching proof-of-concept stage. The first, FUSASEYD, addresses Fusarium Head Blight (FHB), a major fungal disease in winter wheat. Using RGB field images and a deep learning instance segmentation model (YOLOv11), the application detects and quantifies FHB symptoms on wheat ears, offering an automated alternative to time-consuming expert visual scoring. GEVES has developed both a PC interface and a smartphone application to visualise model predictions in the field. Validation in French registration trials is planned for the 2026 campaign. A companion article by V. Cadot et al. is currently under review in the Journal of Experimental Botany special issue on Plant Phenomics & Enviromics Across Scales. The second, COYL (Counting Orange and Yellow Larvae), tackles a practical challenge faced by breeders, and rapidly counting wheat blossom midge larvae, both Sitodiplosis mosellana and Contarinia tritici, to characterise variety susceptibility. Using smartphone RGB images and YOLOv-based object detection, the best-performing model achieved high accuracy and successfully distinguished between the two visually similar species. An online counting application has been developed, currently accessible to Walloon Agricultural Research Centre members. The labelled COYL-1 dataset is publicly available at https://doi.org/10.5281/zenodo.19402333 for community use. A companion article by Antoine Deryck et al. is under submission at Plant Phenomics Journal.
Why it matches plant phenotyping methodsRGB画像と深層学習によりコムギ穂の病徴を検出・定量する手法の開発とセンサー／画像手法の検証が中心であり、植物病害表現型の取得に該当する。
abstractPHENET's Use Case 1 on plant health is validating sensors and imaging methods for the assessment of wheat ear diseases
The architecture of the root system is a primary factor in determining rootstock performance, affecting water and nutrient uptake, biomass accumulation, and overall vigor. However, direct root phenotyping is destructive, labor-intensive, and difficult to do routinely in breeding programs. The present study investigated early root morphological variation among developed interspecific tomato rootstock candidates (Solanum lycopersicum x S. habrochaites). The ability of linear regression and machine learning models to predict root traits from easily measured plant growth parameters was assessed. Nineteen interspecific hybrid rootstock candidates, two commercial rootstocks, and one scion were grown under optimal greenhouse conditions and evaluated at 0, 10, 20, and 30 days after planting. Root length, root surface area, root diameter, and root volume were determined by digital image analysis. In contrast, genotype, plant length, and stem diameter were used as input variables. Significant genotype x sampling date effects were observed for most morphological and biomass traits, indicating dynamic changes in root and shoot development during the first 30 days of growth. The rootstock candidates RSH-17 and RSH-6 generally showed relatively higher root length, surface area, root volume, and biomass accumulation than the commercial rootstocks and scion. XGBoost and OLR were the best predictive models, with R 2 values as high as 0.95 for root length, surface area, and volume. Root diameter was predicted less accurately than root length, surface area, and volume, suggesting that it might be a more independent or less variable root trait during early development. Overall, results suggest that vigor-related traits can serve as useful proxies for estimating major root architectural traits in early-stage tomato rootstock selection. Both XGBoost and OLR performed well, suggesting that root and shoot development were highly coordinated under optimal (non-stress) conditions. Hence, predictive modeling may help prioritize promising rootstock candidates before destructive root analysis. However, more validation under stress conditions and for longer periods of development is needed to determine the greater applicability of these models.
Why it matches plant phenotyping methods根系形態形質をデジタル画像解析で取得し、線形回帰・機械学習による非破壊予測モデルを評価・比較しており、植物フェノタイピング手法が研究の中心である。
abstractThe ability of linear regression and machine learning models to predict root traits from easily measured plant growth parameters was assessed.
Accurate retrieval of leaf area index (LAI) is vital for crop monitoring and genetic breeding. Although multi-modal unmanned aerial vehicle (UAV) remote sensing has advanced LAI estimation, conventional empirical models often overfit on small breeding populations and cannot disentangle the true causal effects of genetic backgrounds from confounding factors within a statistically rigorous framework. This study presents a robust framework for plot-scale maize LAI estimation across 800 breeding plots from four genetic subgroups: doubled haploid (DH), mixed, temperate (TEM), and tropical/subtropical (TST). From UAV RGB and multispectral imagery acquired at three phenological stages, we extracted 76 multi-modal features comprising point-cloud structural metrics, spectral vegetation indices, and texture features. Following a dual-criterion mutual information and multicollinearity filter, four algorithms, including traditional tree-ensembles and the Tabular Prior-data Fitted Network (TabPFN), were evaluated using nested validation. TabPFN achieved superior generalization performance, yielding a mean test R2 of 0.778 ± 0.031, an RMSE of 0.264 ± 0.024, and an MAE of 0.203 ± 0.022, significantly outperforming tree-ensemble models (p < 0.01). Across growth stages, retrieval accuracy peaked at the expanded bell-mouth stage (R2 = 0.802) and successfully captured the unimodal trajectory of canopy development. SHAP-based attribution showed that spectral indices contributed most to the predictions (55.9%), followed by canopy texture (26.3%), with the spatial heterogeneity metric Tex_Entropy being the most influential single feature (22.8%). When embedded as the nuisance estimator within a Double Machine Learning framework for causal inference, TabPFN confirmed that, relative to the TEM subgroup, only the DH genetic background exerted a consistent and significant negative causal effect on LAI (ATE = −0.070, p = 0.030). These results establish TabPFN as a reliable and extensible tool for non-invasive, high-throughput phenotyping in precision breeding.
Why it matches plant phenotyping methodsUAVマルチモーダル画像からトウモロコシLAIを推定する計算・画像解析ワークフローを開発・比較検証し、高スループット表現型解析への応用を示しているため、方法が中心的である。
abstractThis study presents a robust framework for plot-scale maize LAI estimation across 800 breeding plots from four genetic subgroups
Pearl millet is an important crop in arid regions, but its yield is reduced by foliar diseases like Downy Mildew and Rust. Traditional and deep learning methods struggle with accurate lesion detection, severity estimation, and robustness under complex field conditions, and often lack interpretability for practical agricultural deployment. To address these challenges, this study proposes the Adaptive Severity-Aware Swin Attention Network (ASA-SAN), an integrated framework designed for disease segmentation, classification, and severity estimation in pearl millet leaves. The proposed architecture combines a Swin Transformer encoder for hierarchical feature extraction with a ResUNet++ decoder for accurate lesion segmentation. This is further enhanced using Adaptive Channel Attention to improve feature discrimination and a dual-stream classification network to jointly capture local lesion characteristics and global contextual information. Additionally, an Adaptive Disease Severity Index (ADSI) is introduced to quantitatively assess disease progression based on lesion area ratio, color degradation, edge irregularity, and texture variations. Experimental evaluations conducted on a pearl millet leaf dataset demonstrate that the proposed method achieves a Dice score of 97.8%, IoU of 95.6%, classification accuracy of 98.3%, and F1-score of 98.2%, outperforming several state-of-the-art methods. Furthermore, Grad-CAM visualizations enhance model interpretability by highlighting disease-relevant regions. Overall, the ASA-SAN framework provides a robust, interpretable, and severity-aware solution for automated pearl millet disease analysis, enabling early detection and supporting precision agriculture practices for improved crop protection and yield optimization.
Why it matches plant phenotyping methods真珠粟葉の病斑を画像から分割・分類し、病害重症度を定量推定する手法を中心に開発・評価しているため、植物表現型計測手法として含める。
abstractAdditionally, an Adaptive Disease Severity Index (ADSI) is introduced to quantitatively assess disease progression based on lesion area ratio, color degradation, edge irregularity, and texture variations.
Reproduction assets foundThe paper's phenotyping inputs are drawn from a public, open-access image dataset: the Pearl Millet Leaf Disease dataset (Version 2) hosted on Roboflow Universe, containing annotated images of Downy Mildew, Rust, and healthy pearl millet leaves. This is a paper-specific, publicly available asset directly used for the作者Dataset · publicThe dataset used in this research was taken from the publicly available open-access Pearl
Millet Leaf Disease dataset hosted on Roboflow Universe, which has images of Downy
Mildew, Rust and healthy pearl millet leaves annotated publicly available [26]. To ensure
experimental consistency and reproducibility, all experiments were conducted with Version
2 of the open access dataset.Open asset ↗Roboflow Universepdf-raw-page:10 lines:1-28Code / dataset availability confirmedOpenAlex · Crossref · checked 15 Sept 2026
LeafScans-Orchard is a curated, multi-year RGB image dataset of orchard plant leaves designed to support research in computer vision, machine learning, and plant phenotyping. The dataset comprises 9708 high-quality leaf scans acquired during collection campaigns conducted between 2015 and 2025, covering seven orchard crop species: apple, pear, sweet cherry, sour cherry, plum, peach, and apricot. In total, the dataset includes 67 cultivar labels. All samples were acquired using flatbed scanning under controlled conditions on a uniform background, ensuring high visual consistency and minimal background variability. The original scans were captured at 1200 dpi and subsequently converted into a public release format at 300 dpi, stored as lossless TIFF images to preserve morphological and textural details. Each image corresponds to a single leaf and is organized in a hierarchical directory structure by species, cultivar, and acquisition year, accompanied by image-level metadata and aggregated species–cultivar–year counts. LeafScans-Orchard is suitable for plant species classification, cultivar recognition, leaf morphology analysis, texture analysis, and general visual feature extraction. In addition to the main release, a representative subset of 300 original 1200 dpi scans is provided to support high-resolution analyses. The dataset is particularly suited for fine-grained classification, morphology-driven analysis, and methodological studies under controlled imaging conditions.
Why it matches plant phenotyping methods果樹葉のRGBスキャン画像を収録した公開データセットで、植物フェノタイピングおよび葉形態解析を目的とする。標準化された画像取得と再利用可能なデータ構成が中心であり、フェノタイピング用データセットとして適格。
abstractLeafScans-Orchard is a curated, multi-year RGB image dataset of orchard plant leaves designed to support research in computer vision, machine learning, and plant phenotyping.
Reproduction assets foundThe paper's core asset is the LeafScans-Orchard dataset itself (9708 RGB leaf scans, 300 dpi TIFF release plus 1200 dpi subset, image-level metadata and summary counts), openly deposited on Zenodo with an explicit DOI and CC BY 4.0 license. This is a paper-specific, public, actionable phenotyping image dataset. No codeDataset · publicthe published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The dataset described in this article is openly available in Zenodo
as LeafScans-Orchard Dataset (v1.0.0) at https://doi.org/10.5281/zenodo.20187966 (accessed on
10 May 2026). The repository includes the 300 dpi image release, the 1200 dpi high-resolution subset,
image-level metadata, aggregated species–cultivar–year counts, and supporting documentation. The
complete archive of original 1200 dpi scans is retained locally by the authors but is not included in
the current pubOpen asset ↗Zenodo · 10.5281/zenodo.20187966pdf-raw-page:12 lines:1-46Plant phenotyping relevance match · UnverifiedCrossref · checked 14 Sept 2026
Published22 Jun 2026Revista edUCA - Revista Multidisciplinar da Faculdade Católica PaulistaCited by 0 · OpenAlex ↗
The development of Low-Cost Phenotyping Platforms Supported by Generative Artificial Intelligence (AI) is part of a recently launched research initiative at Embrapa Vegetables in Brasília, Federal District, aimed at creating a National Platform for Adaptation to Climate Change Applied to Family Farming (Clima AF). Through Prompt Engineering and Command Chaining, this stage was designed for the visual assessment of physiological disorders in sweet potato (Ipomoea batatas) tuberous roots in the context of the Climate Emergency. The pipeline consists of four stages: 1 - Definition of an expert persona; 2 - Phenological contextualization and critical root filling period; 3 - Visual anatomical phenotyping; and 4 - Synthesis of the physiological disorders found, with a focus on heat stress. The methodology is available as open access following FAIR principles. The analysis is conducted using minimal information, such as photos that can be taken with everyday devices like smartphones and information about the harvest season. Because it is available as open access, it democratizes information and contributes to achieving climate justice for a socioeconomically vulnerable audience (family farmers).
Why it matches plant phenotyping methods生成AIとプロンプト連鎖を用い、スマートフォン画像からサツマイモ塊根の生理障害を視覚的に評価する低コスト表現型解析プラットフォームの開発であり、植物状態の取得・抽出法が中心である。
abstractThe development of Low-Cost Phenotyping Platforms Supported by Generative Artificial Intelligence (AI)
Sapodilla (Manilkara zapota), or chickoo, is a key tropical fruit, very popular in India, Mexico, and Thailand, as it is nutritionally and economically valuable. Nonetheless, the production of sapodillas is often affected by several diseases, which reduce fruit quality and quantity. The dataset used in this paper is a sapodilla fruit image dataset, comprising 1,518 images, gathered in the field under the practicing conditions on 18 February 2025, 22 February 2025, in Rahu village, Pune district, Maharashtra, India, with the use of smartphone cameras. The data is sorted into four categories, namely: Anthracnose, Bacterial rot, Healthy, and Sap bleeding. The photographs were taken in different backgrounds and in different lighting conditions to represent real-life cultivation conditions. The data is expected to be useful in machine learning-based plant disease detection, classification, and analysis, and spur the creation of intelligent and sustainable agricultural systems.
Why it matches plant phenotyping methodsサポディラ果実の病害・健全状態を画像で記録したデータセット自体が中心で、植物病害の画像ベース表現型解析に利用できる。
titleA curated image dataset for sapodilla fruit (Manilkara zapota) disease and fruit quality analysis.
Reproduction assets found保存済みの本文根拠を更新済みルールで再検証し、公開資産1件を確認しました。Dataset · publicData identification number: Version: V1, Doi: 10.17632/xbzd2fjd3p.1
Direct URL to data: https://data.mendeley.com/datasets/xbzd2fjd3p/1Open asset ↗10.17632/xbzd2fjd3p.1html-lines:1-114Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · Europe PMC · checked 5 Sept 2026
Quantitative phenotyping of pepper seedlings is important for greenhouse plug tray seedling cultivation, but it remains constrained by inefficient manual monitoring, complex greenhouse backgrounds, and growth-stage-dependent discrepancies between two-dimensional image traits and actual leaf biomass. In this study, a cascaded vision framework with stage-specific morphological correction was developed for nondestructive seedling phenotyping. The framework integrated Visual Dynamic Momentum YOLO (VDM-YOLO) for individual seedling localization and growth-stage recognition, Variance Guided Strip Ghost Gated UNet (VSG-UNet) for lightweight, high-resolution leaf segmentation, and a stage-aware correction model for leaf dry biomass estimation. In performance evaluation, VDM-YOLO achieved a mean average precision at an intersection over union threshold of 0.5 (mAP0.5) of 89.27%, improving mAP0.5 by 1.82 percentage points over YOLOv12. VSG-UNet achieved a mean intersection over union (mIoU) of 83.9% and a Dice coefficient of 81.8%, while reducing floating point operations (FLOPs) and parameters by 44.2% and 61.2%, respectively, compared with U-Net. After stage-aware calibration, the coefficient of determination (R2) between segmented area and leaf dry weight increased from 0.764 to 0.813, and the root mean square error (RMSE) decreased from 0.0210 g to 0.0190 g. These results demonstrated that the proposed framework provided a proof of concept approach based on RGB images for the nondestructive assessment of leaf area and leaf dry biomass in pepper seedlings under restricted experimental conditions.
Why it matches plant phenotyping methodsRGB画像による葉の検出・セグメンテーションと、葉面積から葉乾燥バイオマスを推定する手法を開発・評価しており、植物表現型取得が研究の中心である。
abstracta cascaded vision framework with stage-specific morphological correction was developed for nondestructive seedling phenotyping.
Wheat is a major staple crop, and storage mold growth poses a severe threat to grain safety and quality stability. Natural mold development in stored wheat exhibits subtle, localized, and highly heterogeneous characteristics. Existing unimodal methods and global fusion approaches generally suffer from insufficient local feature sensitivity, hindering fine-grained mold severity grading. To address this limitation, we propose a Mask-Guided Fine-Grained Fusion Network, a weakly supervised framework based on local RGB-HSI fusion. This framework employs a dynamic parallel A/B experimental design to construct time-matched proxy labels via weakly supervised learning. A standardized preprocessing pipeline including single-kernel extraction, foreground segmentation, and cross-modal registration is established to resolve RGB-HSI spatial misalignment, ensuring physical-level spatial consistency of multimodal features. The model incorporates a Foreground-Aware Spectral Recalibration (FASR) module to suppress background noise, a Mask-Guided Dilated Cross-modal Local Attention (MDCLA) mechanism to establish fine-grained local mappings between RGB visual phenotypes and hyperspectral responses, and a sample-level adaptive fusion strategy to dynamically weight features by modal reliability, enhancing representation of complex samples across all mold stages. Experiments show that the Mask-Guided Fine-Grained Fusion Network achieves 0.9689 classification accuracy, 0.9698 Macro-F1 score, and 0.0593 Mean Absolute Error (MAE), significantly outperforming state-of-the-art unimodal deep models and global attention fusion baselines. This work provides a proof-of-principle framework for fine-grained non-destructive mold risk assessment in stored wheat.
Why it matches plant phenotyping methodsRGB-HSI融合と弱教師あり学習により、保存小麦粒のカビ状態・重症度を推定する手法が研究の中心であり、植物器官の病害状態を直接評価している。
abstractwe propose a Mask-Guided Fine-Grained Fusion Network, a weakly supervised framework based on local RGB-HSI fusion.
Rice, a staple food for more than half of the global population, faces significant yield and quality losses due to various diseases and environmental stresses. This paper presents a new Simplicial Finite-Element-Informed Neural Network based on the Musical Chairs Optimisation Algorithm (SFEINN-MCOA) to detect rice leaf disease precisely and efficiently. The first step involves the acquisition of the RGB image from the Rice Leaf Disease dataset. The Iterative Robust Peak-Aware Guided Filter (IRPAGF) is used to remove noise and enhance contrast, thus improving the image. The Graph-Based Soft-Balanced Fuzzy Clustering (GSBFC) method is used to separate diseased regions and then analyse them. The Self-Distillated Masked Autoencoder (SMA) is used to perform feature extraction and capture important attributes of leaves. The SFEINN classifies data under the category of healthy and diseased leaves, and the MCOA optimises the model parameters to achieve maximum accuracy and minimum error. Experimental findings reveal that the SFEINN-MCOA model has an accuracy of 99.9% and an F1-score of 98.9%, which is better and stronger. This smart system offers a secure, automatic, and effective system for early disease identification of rice and helps farmers to enhance crop health and yield sustainability.
Why it matches plant phenotyping methodsイネ葉の病斑領域を画像から抽出し、健全・罹病状態を自動判定する画像解析フレームワークが研究の中心であり、植物病害状態の表現型推定に該当する。
abstractThis paper presents a new Simplicial Finite-Element-Informed Neural Network based on the Musical Chairs Optimisation Algorithm (SFEINN-MCOA) to detect rice leaf disease precisely and efficiently.
Saudi Arabia is expanding its domestic coffee sector under Vision 2030, yet coffee farming remains vulnerable to leaf diseases and pest damage. Image-based artificial intelligence studies conducted under Saudi field conditions remain limited, particularly in relation to assessing image-based visible disease severity. This study designs a hierarchical deep learning framework for screening coffee leaf diseases using field-collected images of Saudi coffee leaves. Three tasks were addressed: binary health status classification, four-class disease or pest damage identification, and binary visible severity classification. A dataset of 550 RGB images was collected from Al-Dayer Governorate, Jazan, under natural field conditions. ResNet50, DenseNet121, and EfficientNet-B0 were evaluated via transfer learning in two phases: a Saudi-only phase and an integrated phase that combined Saudi data with selected JMuBEN and JMuBEN2 samples. In the Saudi-only phase, ResNet50 achieved 96.47% accuracy for binary classification, while DenseNet121 achieved 68.66% and 78.12% for disease and visible severity classification, respectively. In the integrated phase, performance improved to 99.74%, 97.76%, and 97.37%. These integrated-phase results are interpreted as evidence that dataset expansion and increased visual diversity can improve model performance, rather than as definitive estimates of field deployment performance. The results show that binary classification is feasible under limited local data, whereas fine-grained disease classification is more constrained by dataset size and class imbalance. Grad-CAM visualizations were used to support qualitative interpretability and should not be interpreted as biological validation of disease localization. The framework is positioned as a decision-support screening approach that requires further expert-validated, multi-farm, and multi-season evaluation before deployment.
Why it matches plant phenotyping methodsコーヒー葉の画像から健康状態、病害・害虫損傷、可視的重症度を推定する階層的深層学習フレームワークが研究の中心であり、植物病害状態の画像ベース表現型推定に該当する。
abstractThis study designs a hierarchical deep learning framework for screening coffee leaf diseases using field-collected images of Saudi coffee leaves.
Three-dimensional (3D) reconstruction based on structure from motion and multi-view stereo (SfM-MVS) is increasingly used in plant phenotyping, but its performance is influenced by crop architecture, viewpoint configuration, and image preprocessing. For compact crops such as peanut, dense branching and severe within-canopy occlusion make reliable reconstruction challenging. This study evaluated the effects of reconstruction pipeline, angular interval, and image preprocessing on 3D reconstruction of peanut plants under controlled rotary imaging. A total of 10,800 RGB images from 30 plants were used to compare representative implementations of incremental and global SfM-MVS pipelines in terms of geometric quality, phenotypic accuracy, and processing efficiency. At the 1° baseline, the tested global pipeline implementation reduced the root mean square reprojection error (RMSRE), point-density coefficient of variation (CV), vertical root mean square error (VRMSE), and the 95th percentile of the absolute point-cloud distance values (P95) by 15.05%, 14.08%, 39.87%, and 33.33%, respectively, and increased average phenotypic accuracy from 96.08% to 97.37%, compared with the tested incremental pipeline implementation. In contrast, the tested incremental implementation showed a lower voxel void ratio and shorter processing time. In both pipelines, increasing the angular interval reduced processing time but also reduced geometric stability, internal voxel filling, and phenotypic accuracy. In the present dataset, angular intervals of 3°–5° provided a favourable balance between reconstruction accuracy and efficiency. Cropping reduced peripheral redundancy, whereas cropping combined with background removal produced the best overall results, with the lowest reprojection error and the highest phenotypic accuracy. These results provide practical guidance for selecting reconstruction pipeline, viewpoint configuration, and preprocessing strategy in close-range indoor 3D phenotyping of peanut plants and crops with similar canopy architectures.
Why it matches plant phenotyping methods落花生の3D表現型取得について、SfM-MVSパイプライン、視点間隔、画像前処理を比較・検証しており、フェノタイピング手法が研究の中心である。
abstractThis study evaluated the effects of reconstruction pipeline, angular interval, and image preprocessing on 3D reconstruction of peanut plants under controlled rotary imaging.
The canopy closure stage is a critical phase of rice ( Oryza sativa L.) development that influences canopy structure and final grain yield. Accurate and continuous monitoring of canopy closure dynamics is therefore essential for variety screening and cultivation optimization. This study combines unmanned aerial vehicle (UAV) remote sensing technology with deep learning-based semantic segmentation to establish an efficient framework for quantifying rice canopy closure dynamics. UAV RGB images were acquired for 198 hybrid rice varieties during early growth stages and used to build a canopy segmentation dataset. Three semantic segmentation models, i.e., DeepLabv3+, U-Net, and PSPNet, were systematically evaluated. Results show that DeepLabv3+ performed the best and enabled precise extraction of rice canopy features, obtaining a mean intersection over union (mIoU) of 0.86. Based on the extracted canopy coverage, the Gompertz model was utilized to characterize temporal canopy closure trajectories for all varieties, achieving an average R 2 of 0.978. Subsequently, five key dynamic indicators were derived, including canopy closure limit value ( K ), initial growth coefficient ( a ), growth rate coefficient ( b ), maximum instantaneous growth rate ( MGR ), and days to maximum growth rate ( Tm ). K-means clustering analysis was performed on these indicators to categorize all rice varieties into three clusters, disclosing pronounced differences in early-stage canopy development characteristics. Correlation analysis further demonstrated that canopy closure dynamics were closely associated with grain yield. Overall, while acknowledging the limitations of a single-season and single-site dataset, this study provides a scalable and objective framework for quantifying rice canopy closure dynamics, offering valuable support for variety selection, cultivation optimization, and high-yield rice production.
Why it matches plant phenotyping methodsUAV画像と深層学習セマンティックセグメンテーションを用いてイネのキャノピー閉鎖動態を定量化する手法を構築・評価しており、植物形質抽出が研究の中心である。
abstractThis study combines unmanned aerial vehicle (UAV) remote sensing technology with deep learning-based semantic segmentation to establish an efficient framework for quantifying rice canopy closure dynamics.
Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · checked 5 Sept 2026
Plant diseases result in the estimated loss of 20–40% of the world’s crop production annually, amounting to more than $220 billion in economic losses and threatening food security for a rapidly expanding world population. While the conventional methods for detecting plant diseases rely on visual inspection of the symptoms, they are resource-consuming. For effective plant disease detection at a pre-mature stage, hyperspectral imaging (HSI) represents a paradigm shift in technology. It can be used to obtain subtle spectral signatures outside the visible spectrum, which enables pre-symptomatic and highly specific plant disease diagnosis. Concurrently, deep learning (DL) has become the prevalent analytical paradigm for decoding the complex and high-dimensional data that HSI produces. This paper covers a comprehensive narrative review of the intersection of these two transformative technologies from 2008 to 2026. We first set out the biological and physical principles by which HSI is uniquely suited to detecting plant–pathogen interactions in the absence of visible symptoms. We then present a detailed taxonomy of deep learning architectures for Vision Imaging and HSI data, ranging from basic 1D and 3D convolutional neural networks (CNNs) to hybrid models with attention mechanisms and, most recently, vision transformers, which have achieved greater robustness to real-world conditions. There is currently a major and consistent “lab-to-field” performance gap. A critical analysis of various studies reveals a persistent and significant performance gap between models that perform well on controlled lab datasets (ranging from 95 to 99%) and field-collected data (typically 70–85%). This paper also addresses the practical gap of environmental variability, image noise, and the domain gap between the controlled environment and the real dataset. Finally, this review concludes by providing strategic research recommendations and a roadmap, highlighting that the future of the field is contingent upon not only architectural innovation but also a holistic approach, with robustness, scalability, affordability, and interpretability as the main focus to bring the proven potential of HSI-DL systems from the lab to the field, ultimately contributing to global food security.
Why it matches plant phenotyping methods植物病害の症状・状態をハイパースペクトル画像と深層学習で推定する手法を中心に扱うレビューであり、植物フェノタイピング手法レビューに該当する。
abstractThis paper covers a comprehensive narrative review of the intersection of these two transformative technologies from 2008 to 2026.
The Vines-DB dataset contains 1,218 original high-resolution RGB images of seven ornamental vine species collected under field conditions at the Utah Agricultural Experiment Station's Greenville Research Farm in Logan, Utah, USA. The dataset was generated from 168 individual vine plants that were transplanted in 2022 and photographed repeatedly across multiple months during the 2023 and 2024 growing seasons (July-October). Images were captured with an iPhone 16 Pro equipped with a 48 MP camera between 10:00 AM and 12:00 PM under daylight. Vines were grown on 1.2m x 2.4m trellises and photographed from a distance of 1m against black or white Styrofoam backdrops to improve contrast and reduce background noise. The dataset includes Akebia quinata, Campsis radicans, Hydrangea anomala petiolaris, Lonicera x heckrottii, Campsis x tagliabuana 'Madame Galen', Parthenocissus quinquefolia, and Wisteria floribunda. All original images were manually annotated in Roboflow by trained annotators to produce polygon-based instance segmentation masks for eight classes, including seven species and background. After preprocessing and data augmentation, the working dataset was expanded to 2,307 images for model development and evaluation. The augmented dataset was divided into 2,019 training images, 192 validation images, and 96 test images using stratified sampling to maintain balanced representation. Vines-DB supports the development and evaluation of deep learning models for multi-class instance segmentation in precision horticulture and urban ecology. The dataset enables applications such as automated canopy cover estimation, species identification, and scalable field phenotyping. In addition, repeated monthly imaging of the plants captures temporal variation in canopy development and plant appearance, increasing the dataset's utility for segmentation benchmarking under realistic field conditions.
Why it matches plant phenotyping methods植物のRGB画像とポリゴン注釈から成るデータセットを構築し、セグメンテーション評価およびキャノピー被覆推定などの植物フェノタイピングを支援することが中心であるため。
abstractVines-DB supports the development and evaluation of deep learning models for multi-class instance segmentation in precision horticulture and urban ecology.
Reproduction assets foundThe paper's core asset is the Vines-DB RGB image dataset with instance segmentation annotations, publicly deposited on OSF with an explicit DOI and URL matching an allowed URL.Dataset · publicData accessibility Repository name: Vines-DB
Data identification number: 10.17605/OSF.IO/YJHCK
Direct URL to data: https://osf.io/yjhck/overviewOpen asset ↗OSF · 10.17605/OSF.IO/YJHCKpdf-page:2 lines:1-49Code / dataset availability confirmedCrossref · Europe PMC · checked 5 Sept 2026
Early and accurate disease detection is important for increasing the agricultural output, decreasing the financial costs, and ensuring food security. Traditional diagnostic procedures take much time and effort, involve the necessity of having deep expertise, and are not always suitable for large scale farming disease detection. For this purpose, the current research suggests developing an explainable lightweight CNN-based model for crop disease identification based on RGB leaf images. The model utilizes several innovative architectural solutions such as depth-wise separable convolution, SE blocks, skip connections, and guided attention-based feature learning that allow enhancing the effectiveness of features extraction and decreasing computation load. Moreover, Grad-CAM is used to visualize affected areas on a map and thus increase the interpretability of the model. The suggested solution was implemented and tested on the PlantVillage dataset containing 54,305 images for 38 crop diseases out of 14 crops. The results show that the training, validation, and testing accuracies equal 97.6%, 88.3%, and 97.63%, correspondingly, along with the Macro-F1 measure of 0.867 and Micro-ROC-AUC equal to 0.99. A comparative study reveals that the presented model performs comparably well in terms of classification with lightweight structure and built-in interpretability capabilities to be applied in the mobile and edge-enabled agriculture environment. The results show that the presented approach is capable of being used as an effective and interpretable tool for diagnosing plant diseases in real-time.
Why it matches plant phenotyping methods葉画像から植物病害を推定するCNN手法の開発・評価が研究の中心であり、植物の病害状態を直接推定するため、植物フェノタイピング手法として含める。
abstractthe current research suggests developing an explainable lightweight CNN-based model for crop disease identification based on RGB leaf images.
Reproduction assets foundThe paper's plant-phenotyping input is the public PlantVillage leaf-image dataset (54,305 RGB images, 38 crop-disease classes), explicitly declared in the Data availability statement with a Kaggle URL. No author code, trained model, or checkpoint is deposited.Dataset · publicThe data set analyzed during current study are available in https://www.kaggle.com/datasets/emmarex/plantdisease.Open asset ↗Kaggle · emmarex/plantdiseaselines:366-390Plant phenotyping relevance match · UnverifiedCrossref · checked 15 Sept 2026
Oil palm is an essential commodity for the economy; however, basal stem rot caused by Ganoderma boninense poses a significant threat to plantation productivity and long-term vitality. It highlights the importance of early detection of stem disease to facilitate timely intervention and minimize potential economic losses. This study presents an image-based approach to diagnosing oil palm stem maladies, leveraging handcrafted color and texture features within a supervised machine learning framework. The dataset contained 525 images of oil palm stems, of which 205 depicted healthy specimens, and 320 depicted diseased ones. These were captured within their natural environment. Color features were derived by analyzing color moments within the HSV color space, while texture features were extracted from the Grey-Level Co-occurrence Matrix (GLCM). The extracted features were classified employing an Artificial Neural Network (ANN) and were subsequently contrasted with classifiers including Decision Tree, K-Nearest Neighbors, Naive Bayes, and Support Vector Machine. Model performance was evaluated using k-fold cross-validation with k = 5 and k = 10 to ensure the consistency and reliability of the assessment. The experimental results demonstrated that the highest accuracy of 97.52% was achieved when the ANN model was used to classify the integrated color and texture features. The innovative aspect of this research resides in demonstrating that handcrafted features integrated with artificial neural networks can attain high detection accuracy in scenarios with limited data, providing a viable alternative to data-intensive deep learning techniques. This method facilitates a dependable, computer vision-driven early detection system for oil palm stem diseases, thereby promoting sustainable plantation management.
Why it matches plant phenotyping methods油ヤシ幹の病徴を画像から色・テクスチャ特徴として抽出し、分類器で病害状態を推定する方法が研究の中心であり、交差検証による性能評価も行っているため、植物表現型計測手法として含める。
abstractThis study presents an image-based approach to diagnosing oil palm stem maladies, leveraging handcrafted color and texture features within a supervised machine learning framework.
In-season estimation of wheat grain yield potential is critical for crop management and advancing breeding efforts. Spike and spikelet counts serve as key indicators directly linked to yield potential, yet their assessment still relies on manual counting which is both labor-intensive and error-prone. High-resolution digital (RGB) imagery combined with deep learning-based object detection methods has substantially advanced automatic wheat spike detection and counting. However, precise spikelet-level phenotyping remains largely underexplored. This study evaluates two recent YOLO variants, YOLOv11 and YOLOv12, for wheat spike and spikelet detection and counting using oriented bounding boxes (OBB), and introduces a new large-scale benchmark dataset comprising 48,521 spike and 60,404 spikelet instances with OBB annotations. For spike detection, the pre-trained YOLOv11 achieved superior accuracy (mAP@0.5 = 95.8%, Pearson r = 0.993) with shorter training and inference times compared to YOLOv12. For spikelet detection, the non-pretrained YOLOv11 demonstrated higher accuracy (mAP@0.5 = 99.0%), while counting performance was comparable across models. These results establish OBB-based YOLO detection as a robust and scalable approach for AI-driven wheat phenotyping.
Why it matches plant phenotyping methods小麦の穂・小穂という収量関連形質の画像ベース検出・計数手法を比較評価し、大規模ベンチマークデータセットも構築しているため、フェノタイピング手法が中心である。
abstractThis study evaluates two recent YOLO variants, YOLOv11 and YOLOv12, for wheat spike and spikelet detection and counting using oriented bounding boxes (OBB), and introduces a new large-scale benchmark dataset comprising 48,521 spike and 60,404 spikelet instances with OBB annotations.
Reproduction assets foundThe paper openly states its supporting data (spike/spikelet imagery with OBB annotations) is available on Zenodo, and the underlying models are deployed on the authors' public WheatAI cloud platform.Dataset · publicData availability
The data supporting the findings of this study are openly available at: https://doi.org/10.5281/zenodo.20215489 .Open asset ↗zenodo · 10.5281/zenodo.20215489lines:219-266Plant phenotyping relevance match · UnverifiedCrossref · checked 15 Sept 2026
Published15 Jun 2026Journal of Electrical Engineering and InformaticsCited by 0 · OpenAlex ↗
This study aims to develop an edge-based Internet of Things (IoT) system for automatic plant leaf condition classification using classical digital image processing on a Raspberry Pi. The proposed system classifies leaf conditions into healthy, diseased, and pest-attacked categories while providing real-time remote monitoring through a Telegram Bot. The system employs a Raspberry Pi as the edge computing device and a Raspberry Pi Camera for image acquisition. Images are processed locally using OpenCV through RGB-to-HSV color space conversion, thresholding, edge detection, and contour analysis. System performance was evaluated using 15 test samples for each image acquisition distance (15 cm, 30 cm, 45 cm, and 60 cm). Experimental results achieved detection accuracies of 100% at image acquisition distances of 15 cm and 30 cm, while the accuracy decreased to 87% at 45 cm. At 60 cm, the system failed to detect the target object because insufficient visual information prevented reliable feature extraction. The proposed edge-based IoT system provides an efficient and low-cost solution for real-time plant leaf condition classification. The experimental results indicate that an image acquisition distance of 15-30 cm is optimal for reliable detection under the evaluated experimental conditions. The proposed system integrates Raspberry Pi-based edge computing with lightweight classical digital image processing and Telegram Bot notifications, eliminating the need for computationally intensive deep learning models or cloud-based image processing.
Why it matches plant phenotyping methods植物葉の健全・病害・害虫被害状態を画像から分類する手法と、Raspberry Pi上の実装・性能評価が研究の中心であり、植物状態のフェノタイピング手法に該当する。
abstractThis study aims to develop an edge-based Internet of Things (IoT) system for automatic plant leaf condition classification using classical digital image processing on a Raspberry Pi.
Leaf diseases in banana plants and ornamental crops can significantly reduce productivity and product quality, highlighting the need for accurate early detection methods. This study proposes an image-based classification approach utilizing texture features extracted from the Gray Level Co-occurrence Matrix (GLCM) combined with a Hybrid Stacking model that integrates Random Forest (RF) and Support Vector Machine (SVM). The preprocessing stage involves image resizing and noise reduction, followed by feature extraction using energy, contrast, homogeneity, and correlation parameters. The dataset consists of eight classes of healthy and diseased leaves, collected from both field documentation and secondary sources. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics under a cross-validation scheme. Experimental results show that SVM achieved 89.2% accuracy, RF 88.5%, while the stacking model yielded the best performance with 91.7% accuracy, effectively reducing misclassification among visually similar disease classes. This study demonstrates the effectiveness of combining GLCM features and hybrid stacking models for leaf disease classification, with potential applications in automated plant monitoring systems to support precision agriculture.
Why it matches plant phenotyping methods植物葉の病害状態を画像から分類する手法の開発・評価が研究の中心であり、GLCM特徴量とRF/SVMの性能を交差検証しているため、植物フェノタイピング手法として収録する。
abstractThis study proposes an image-based classification approach utilizing texture features extracted from the Gray Level Co-occurrence Matrix (GLCM) combined with a Hybrid Stacking model that integrates Random Forest (RF) and Support Vector Machine (SVM).
Real-time monitoring of H 2 O 2 in plant tissues is useful for evaluating oxidative changes during postharvest storage, but direct on-site detection in vegetables remains difficult because most assays still require tissue disruption and laboratory instruments. In this study, a dual-signal microneedle biosensor was developed by integrating polydopamine-coated Fe/Zr-MOF nanozyme (PDA@Fe/Zr-MOF) into a gelatin/sodium alginate microneedle patch for H 2 O 2 detection in lettuce. The polydopamine coating improved the peroxidase-like response of Fe/Zr-MOF through •OH generation and also contributed to photothermal conversion under 808 nm near-infrared (NIR) irradiation. After contact with lettuce leaves, the microneedles extracted interstitial fluid and allowed H 2 O 2 -triggered TMB oxidation to be read by both colorimetric imaging and thermal imaging. The two outputs were not independent recognition mechanisms, but they provided mutually supportive information and helped reduce the influence of sample color and environmental fluctuations. The sensor achieved detection limits of 0.42 μM for the colorimetric mode and 0.34 μM for the photothermal mode. During 15 days of storage at 4°C, the sensor tracked H 2 O 2 accumulation in lettuce and showed a clear relationship with spoilage progression. These results indicate that PDA@Fe/Zr-MOF-based microneedle sensing is a feasible approach for monitoring oxidative freshness changes in postharvest vegetables.
Why it matches plant phenotyping methodsレタス組織内H2O2という植物の生理状態を、マイクロニードルとカラー・熱画像で現場測定するセンサーを開発しており、取得手法が研究の中心である。
abstracta dual-signal microneedle biosensor was developed by integrating polydopamine-coated Fe/Zr-MOF nanozyme (PDA@Fe/Zr-MOF) into a gelatin/sodium alginate microneedle patch for H 2 O 2 detection in lettuce.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 5 Sept 2026
Abstract Simultaneous stresses of salinity and drought often coincide during rice-growing seasons in coastal areas due to insufficient water resources and inadequate irrigation infrastructure. Consequently, combined salinity-drought stress poses a major threat to rice production. To investigate the effects of combined salinity-drought stress, a two-season study was conducted utilizing soil media. The first season involved screening 58 rice genotypes, while the second season focused on validating the consistency of response in 20 selected tolerant and susceptible genotypes. These included established tolerant checks (Pokkali and Salumpikit) and susceptible checks (IR 29 and IR 20). Both drought and salinity treatments were given at an electrical conductivity (EC) of 10 dSm⁻¹ and 75% field capacity at the seedling stage. The experimental design was arranged in a modified lattice design in each season, with six blocks and three replications in the first season and two blocks and five replications in the second season. The data collected are leaf symptoms, biomass weight, and shoot length. A number of 330 images captured by a smartphone camera. Machine learning models were employed to predict drought-salinity tolerance criteria. The study revealed that XGBoost model achieved an accuracy of 90.62%. The study identified two genotypes, IR18A1925-SKI-0 and Inpari 30, that exhibited insignificance to Pokkali, based on assessment of shoot length, biomass, and leaf symptoms. These two genotypes were also consistently clustered with Salumpikit. These findings highlight potential of machine learning techniques in predicting rice tolerance to combined salinity-drought stress, with the XGBoost model demonstrating superior predictive capability in this study.
Why it matches plant phenotyping methodsスマートフォン画像から葉症状・バイオマス・草丈などの表現型を取得し、機械学習で複合ストレス耐性を予測する手法の適用が研究の中心である。
titleImage-based Phenotyping and Machine Learning Prediction of Rice Genotypes to Combined Drought-Salinity Stresses
Determining the drivers of ecological stability amid accelerating global environmental change is a critical goal of contemporary ecology. Various candidate drivers have been suggested, with recent attention turning to response diversity—the variation among organism-environment responses. However, despite conceptual interest in response diversity as a driver of stability, there remain few field tests of this relationship. Using multi-species competitive communities of floating aquatic macrophytes as an experimental model for measuring temporal stability and response diversity to nutrient loading, we show that response diversity does not promote temporal stability of total macrophyte cover, but that communities with an uneven distribution of species responses were more resistant to an exogenous shock. To quantify macrophyte composition and growth dynamics from photographic time series of our experimental communities, we developed an open-source, scalable, machine learning workflow ( LeafMosaic ) capable of classifying four species from noisy field data including variable lighting, resolution, and plant morphology. We measured response diversity as the balance of positive and negative biomass growth responses to dissolved nitrate concentration, weighted by species’ relative contributions to biomass, and tested its effect on temporal stability and resistance to an unexpected pulse disturbance (a large typhoon that disrupted our outdoor mesocosms). Response imbalance predicted typhoon resistance, but species asynchrony and mean population stability best predicted community stability, with no direct or indirect effect of species responses. Overall, our results provide new experimental evidence for how the structure of species responses promotes stability, and we aim our LeafMosaic workflow to empower future field experiments using floating macrophytes to study response diversity and ecological stability.
Why it matches plant phenotyping methods浮遊水生植物の写真時系列から種組成と成長動態を抽出する、オープンソースでスケーラブルな機械学習ワークフローを開発しており、植物表現型取得・解析法が中心的です。
abstractTo quantify macrophyte composition and growth dynamics from photographic time series of our experimental communities, we developed an open-source, scalable, machine learning workflow ( LeafMosaic ) capable of classifying four species from noisy field data including variable lighting, resolution, and plant morphology.
Agriculture plays a vital role in the economy of many countries, and crop productivity is highly dependent on plant health.Plant diseases can significantly reduce crop yield and quality if not detected at an early stage.Traditional disease identification methods rely on manual inspection by agricultural experts, which can be time-consuming, expensive, and sometimes inaccurate.Recent advancements in image processing and machine learning have enabled automated systems for plant disease detection.This research presents a plant leaf disease detection system based on a Multi-Class Support Vector Machine (SVM) classifier using color, texture, and shape features extracted from leaf images.The proposed approach captures leaf images, performs preprocessing to remove noise and enhance image quality, segments the infected region, and extracts relevant features.These features are then used to train a Multi-Class SVM model capable of classifying different plant diseases.The combination of color, texture, and shape characteristics improves classification accuracy by providing comprehensive information about disease symptoms present on the leaf surface.Experimental analysis demonstrates that the proposed method can effectively identify multiple plant diseases with high accuracy while reducing the dependency on manual diagnosis.The developed system offers a cost-effective and efficient solution for farmers and agricultural professionals, helping in early disease detection and timely treatment recommendations.The proposed approach contributes to the advancement of smart agriculture and precision farming technologies.
Why it matches plant phenotyping methods葉画像から病斑領域を分割し、色・テクスチャ・形状特徴を抽出して植物病害状態を分類する画像ベースの表現型推定手法が研究の中心であるため。
abstractThis research presents a plant leaf disease detection system based on a Multi-Class Support Vector Machine (SVM) classifier using color, texture, and shape features extracted from leaf images.
Abstract Continuous, high‐frequency monitoring is essential to capture rapid phenological transitions and dynamic crop responses to the environment. However, most phenotyping platforms lack the temporal resolution and automation required for consistent, season‐long trait assessment. This study introduces AGIcam, an open‐source Internet of Things (IoT) camera system for automated and continuous in‐field plant phenotyping and yield prediction. The platform integrates solar‐powered Raspberry Pi units with a modular software stack, comprising Node‐RED, InfluxDB, Grafana, and Microsoft Azure, for automated data acquisition, transfer, and visualization. In the 2022 growing season, 18 AGIcam systems were deployed in spring and winter wheat ( Triticum aestivum ) breeding trials, maintaining an uptime of over 85% while capturing frequent red‐green‐blue and no‐infrared imagery. Time‐series vegetation indices derived from these images were used to predict yield using random forest and long short‐term memory (LSTM) models. The LSTM approach achieved the highest accuracy approximately one week after heading, with mean prediction errors of 3.41% for spring wheat and 1.62% for winter wheat. These results highlight the potential of IoT‐based platforms such as AGIcam to enable real‐time, scalable, and effective phenotyping solutions for data‐driven crop improvement. The presented work provides open‐source resources for the development and time‐series analysis of IoT data for phenotyping and precision agricultural applications.
Why it matches plant phenotyping methods植物フェノタイピング用のIoTカメラ基盤を開発・実証し、画像由来の時系列形質から収量を予測する方法が研究の中心である。
abstractThis study introduces AGIcam, an open‐source Internet of Things (IoT) camera system for automated and continuous in‐field plant phenotyping and yield prediction.
The world suffers from 10–40% loss in crop yields each year because of plant disease. This threat is serious and growing; it threatens food security, rural livelihoods, and agricultural economies. Advances being made through deep learning, computer vision, and mobile technology have presented a unique opportunity to use leaf images to automatically recognize plant disease. Published classification accuracies on benchmark datasets now exceed 97%, which is an important achievement but achieving high accuracy on a benchmark alone does not indicate that traditional methods will work when deployed in the real world: all four stakeholders (i.e., farmers, agronomists, regulatory authorities, and extension agents) must therefore have the ability to understand, and interpret the output of automatically recognized plant diseases in a way that enhances human expertise rather than replacing it. In this chapter, we provide a compendium of technical deep learning architectures and methods related to Explainable Artificial Intelligence (XAI) for plant disease detection, including convolutional networks, residual architectures, dense architectures, transformer networks, and hybrid models. We also systematically evaluate the explainability methods used in both post-hoc and intrinsic explanation and evaluate the applicability of these methods across a variety of imaging modalities used in agriculture, including RGB, multispectral, and hyperspectral. This chapter characterizes major benchmark datasets; discusses major challenges to their deployment, including class imbalance, domain shift, model size reduction, and human–AI trust calibration; then ends with potential new directions for research in areas such as foundation models (FM), causal interpretable models (Explanations), federated learning, and continual learning to build resilience for each evolving pathogen landscape.
Why it matches plant phenotyping methods植物病害を葉画像から自動認識する画像ベースの表現型推定手法と、その説明可能性・データセット・評価課題を体系的に扱うレビューであり、方法論が中心である。
abstractIn this chapter, we provide a compendium of technical deep learning architectures and methods related to Explainable Artificial Intelligence (XAI) for plant disease detection
Despite the widespread use of image-only convolutional models for plant disease diagnosis to provide global food security, image variability at the field level, visually similar symptoms, and stress due to soil nutrients or moisture conditions, can cause loss of accuracy. This paper compares classical machine learning classification models to custom convolutional neural networks and pretrained transfer-learning models for multi-crop and multi-disease classification and also introduces a late-fusion multimodal decision support model that could integrate data about images, soil, and weather. Experiments were conducted on the PlantVillage dataset which had 54,305 RGB images belonging to 38 different crop-condition classes (43,456 training, 10,849 validation, and 10,849 test images) across 14 different crops. The accuracy, macro-precision, macro-recall and macro-F1 score were computed for the five classical classifiers (KNN, Random Forest, Extra Trees, SGD-linear SVM, and SVC-RBF using PCA-reduced features), three custom CNN variants, and four pretrained models (ResNet50, MobileNetV2, GoogleNet, and EfficientNetB7). Of the classical models, SVC with RBF kernel yielded the highest accuracy (83.58%) and MacroF1 (82.96%). In the case of the custom CNN models, the deeper they became and the more dropout the higher accuracy they gained, with CNN-V3 attaining 95.71% accuracy and 95.66% macro-F1. Transfer learning yielded the best results with MobileNetV2 (99.30% accuracy and macro-F1) outperforming ResNet50 (98.60% accuracy and macro-F1) and GoogleNet (97.83% accuracy and 97.46% macro-F1). The findings serve as the basis for the introduction of a multimodal system for soil forecasting integrating a MobileNetV2 image encoder, a residual MLP for soil features, and two dual LSTMs for 48-hour and 168-hour time-series of the weather. The interpretability, modularity and the sufficient resistance towards loss of sensor information for probability-level late fusion with 0.80, 0.10, 0.05 and 0.05 respectively make the framework appropriate for precision-agriculture advisory systems, until it is tested in the field.
Why it matches plant phenotyping methods植物画像から病害状態を推定する分類手法を複数比較し、マルチモーダル意思決定フレームワークを構築・評価しており、病害表現型の取得・推定が中心である。
abstractThis paper compares classical machine learning classification models to custom convolutional neural networks and pretrained transfer-learning models for multi-crop and multi-disease classification
Potato sorting in post-harvest processing relies heavily on manual visual inspection, which is physically demanding, subjective, and insufficiently scalable for modern packing lines. This study investigates the feasibility of a low-cost RGB-based optical inspection system for automated potato quality detection using deep learning-based object detection. A controlled imaging platform was constructed using commodity hardware, and a dataset of 19,805 manually annotated instances across 1361 images was collected from two geographically distinct farm locations in Slovenia. A systematic benchmark of 25 model configurations spanning five YOLO architecture families-YOLOv8, YOLOv9, YOLOv10, YOLOv11, and YOLO26-was conducted across three practical quality classes (Edible, Feed, Rotten) using a strict cross-location evaluation protocol in which models were trained on one location and tested on a completely unseen second location. All models achieved strong in-distribution performance (F1 ≥ 0.906), but showed considerable variation under cross-location conditions, with external F1 ranging from 0.792 to 0.918. The yolo26_l configuration achieved the best cross-location performance (F1 = 0.918, mAP@0.5:0.95 = 0.816, ΔF1 = 0.029), demonstrating that transferable representations are achievable under a standard supervised training protocol. Per-class analysis identified feed detection as the primary generalization bottleneck. The results confirm that affordable RGB-based sorting systems are technically feasible and highlight cross-location evaluation as an essential protocol for assessing real-world deployment readiness.
Why it matches plant phenotyping methodsジャガイモ塊茎の品質・腐敗状態をRGB画像と物体検出で推定する撮像システムを構築し、複数YOLOモデルの交差地域ベンチマークと外部検証を行っており、表現型取得法が中心である。
abstractThis study investigates the feasibility of a low-cost RGB-based optical inspection system for automated potato quality detection using deep learning-based object detection.
First stable release of the RGB and NPQ pixel-wise phenotyping pipeline associated with the manuscript "Image-based biomarkers effectively predict salt and drought stress in dwarf tomatoes (Solanum lycopersicum L.)". This repository includes a Python-based image analysis pipeline for high-throughput plant phenotyping using RGB and chlorophyll fluorescence (NPQ) imaging data. The workflow is designed for pixel-wise extraction and analysis of image-derived traits, with a specific focus on preserving full spatial distributions rather than relying on image-level summary statistics. The pipeline processes RGB images to compute vegetation indices derived from color channel combinations, and NPQ fluorescence images to extract pixel-level chlorophyll fluorescence metrics. Both data types are integrated with experimental metadata through structured indexing files. The analysis framework is organised into two main stages: (i) data pre-processing and structuring into long-format pixel-wise datasets, and (ii) distribution-based statistical analysis of trait variability across treatments and conditions. The latter includes normalised histograms, Jensen–Shannon and Wasserstein distance metrics, and cluster-based permutation testing to identify statistically significant differences between distributions. A minimal example dataset is provided to enable end-to-end testing of the workflow, including image processing, metadata integration, and statistical analysis. An additional archive containing representative example outputs generated from the example dataset is included to illustrate the structure and format of intermediate and final pipeline outputs. To facilitate computational reproducibility, the repository also includes complete derived outputs generated from the full study dataset, including distribution-comparison results (Jensen–Shannon and Wasserstein distances) and cluster analysis outputs for all evaluated RGB and chlorophyll fluorescence traits. These files are provided as supplementary computational products of the workflow and can be used to verify, inspect, and reproduce the analyses described in the associated manuscript.
Why it matches plant phenotyping methodsRGBおよびNPQ画像から植物形質を画素単位で抽出・解析する再利用可能なパイプラインと再現性用データを提供しており、フェノタイピング手法が中心である。
abstractFirst stable release of the RGB and NPQ pixel-wise phenotyping pipeline
First stable release of the RGB and NPQ pixel-wise phenotyping pipeline associated with the manuscript "Image-based biomarkers effectively predict salt and drought stress in dwarf tomatoes (Solanum lycopersicum L.)". This repository includes a Python-based image analysis pipeline for high-throughput plant phenotyping using RGB and chlorophyll fluorescence (NPQ) imaging data. The workflow is designed for pixel-wise extraction and analysis of image-derived traits, with a specific focus on preserving full spatial distributions rather than relying on image-level summary statistics. The pipeline processes RGB images to compute vegetation indices derived from color channel combinations, and NPQ fluorescence images to extract pixel-level chlorophyll fluorescence metrics. Both data types are integrated with experimental metadata through structured indexing files. The analysis framework is organised into two main stages: (i) data pre-processing and structuring into long-format pixel-wise datasets, and (ii) distribution-based statistical analysis of trait variability across treatments and conditions. The latter includes normalised histograms, Jensen–Shannon and Wasserstein distance metrics, and cluster-based permutation testing to identify statistically significant differences between distributions. A minimal example dataset is provided to enable end-to-end testing of the workflow, including image processing, metadata integration, and statistical analysis. An additional archive containing representative example outputs generated from the example dataset is included to illustrate the structure and format of intermediate and final pipeline outputs. To facilitate computational reproducibility, the repository also includes complete derived outputs generated from the full study dataset, including distribution-comparison results (Jensen–Shannon and Wasserstein distances) and cluster analysis outputs for all evaluated RGB and chlorophyll fluorescence traits. These files are provided as supplementary computational products of the workflow and can be used to verify, inspect, and reproduce the analyses described in the associated manuscript.
Why it matches plant phenotyping methodsRGBおよびNPQ画像から植物形質を画素単位で抽出・解析する再利用可能なパイプラインと再現性資料が中心であり、植物フェノタイピング手法に該当する。
abstractPython-based image analysis pipeline for high-throughput plant phenotyping using RGB and chlorophyll fluorescence (NPQ) imaging data.
The integration of artificial intelligence (AI), machine learning (ML), and precision agriculture has created new opportunities for efficient and sustainable crop monitoring. These technologies enable large-scale analysis of agricultural data to assess plant health, optimize resource usage, and support data-driven decision-making. This work presents a machine-learning-based framework for assessing strawberry plant health using RGB imagery collected from unmanned aerial vehicles (UAVs). Unlike traditional object detection approaches, this study adopts a hierarchical classification strategy using convolutional neural networks, including different ResNet and EfficientNet architectures. Individual plant regions are extracted as blobs through a preprocessing pipeline, and these image tiles are used to train stage-wise binary classifiers that progressively distinguish plant health categories. To enhance reliability, model predictions are validated using field-collected ground-truth data including chlorophyll measurements and visual plant health ratings, as well as real-time deployment scenarios, where predictions are made from live UAV video feeds. Geospatial alignment associates image-based predictions with real-world measurements, enabling comprehensive evaluation of model performance. Experimental results showed that ResNet18 achieved 87.75% accuracy with an F1 score of 0.8524 for healthy plant classification and 93.75% accuracy with an F1 score of 0.6115 for unhealthy plant classification. EfficientNet-B0 demonstrated superior performance for moderately healthy and moderately unhealthy categories, achieving accuracies of 66.83% and 75.65%, with F1 scores of 0.6350 and 0.6070, respectively, highlighting the effectiveness of the hierarchical classification framework. This framework demonstrates the practical potential of RGB-based plant health monitoring integrated with geospatial alignment and field-validated measurements, offering a scalable, efficient solution for precision agriculture applications.
Why it matches plant phenotyping methodsRGB画像と深層学習を用いてイチゴ個体の健康状態を推定する分類フレームワークを開発し、地上測定・目視評価・実運用映像で検証しており、植物表現型取得が中心的です。
abstractThis work presents a machine-learning-based framework for assessing strawberry plant health using RGB imagery collected from unmanned aerial vehicles (UAVs).
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · Crossref · checked 6 Sept 2026
Abstract Rapid and accurate quantification of crop biomass using multisource UAV imagery–derived features, such as plant height, vegetation indices, and texture indices demonstrates strong potential for soybean high-throughput phenotyping. The indeterminate growth habit of soybean, alongside extensive nodulation and intense inter-plant competition, necessitates individual plant level (IPL) monitoring to quantify plant-specific nitrogen fixation and competitive vigor. However, most studies aggregate measurements at multi-plants at the plot level, thereby masking these soybean-specific traits. This study aims to develop and evaluate a UAV imagery-based framework for estimating soybean biomass at the IPL, with the objective of characterizing high-resolution spatial variability and supporting high-throughput phenotyping. Regions of interest (ROIs) for IPL data acquisition were defined as rectangular plots based on planting density and were generated from early-stage imagery before canopy overlap occurred. Using these ROIs, structural information (SIs) including plant height (PH) and vegetation fraction (VF)), vegetation indices (VIs), and texture indices (TIs) were derived for each individual plant from RGB and multispectral imagery and organized into sensor-specific feature groups. Recursive feature elimination was applied to select optimal features, which were then used as inputs for machine learning architectures including support vector regression (SVR) with a linear kernel, Random Forest (RF), and XGBoost (XGB). Among them, the SVR model using fused multisource features (SIs + VIs + Tis) achieved the best performance, with R² = 0.88, RMSE = 55.34 g, and rRMSE = 8.50% on an independent test dataset. The results show that: (1) tree-based models, including XGB and RF may suffer from overfitting due to limited sample size and feature redundancy, whereas linear SVR showed better generalization; (2) fusing RGB and multispectral features consistently improved biomass estimation accuracy. Inparticular, near-infrared and red-edge-based indices such as RECI and NDRE, along with VF and PH, were identified as important predictors, while texture indices were not selected as significant features; and (3) the proposed framework enabled spatially explicit IPL biomass mapping and time-series analysis, revealing variability in growth conditions and distinct growth trajectories. The framework provides a reliable solution for IPL soybean biomass estimation with practical potential for UAV-based agricultural decision-making.
Why it matches plant phenotyping methodsUAV画像由来の特徴量と機械学習により個体レベルのダイズ biomass を推定・評価する枠組みが研究の中心であり、植物表現型の取得・抽出手法に該当する。
abstractThis study aims to develop and evaluate a UAV imagery-based framework for estimating soybean biomass at the IPL
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 14 Sept 2026
Cowpea is a crop of great importance worldwide, which is why many heirloom varieties and improved cultivars are explored. Consuming pods and green beans provides vitamins, minerals, and functional components for people with limited access to vegetables. The pods and green beans of these materials have intrinsic characteristics that distinguish them. Therefore, the objective was to adjust machine learning models to identify cowpea from digital images of pods and green beans using artificial intelligence techniques. Digital images of four heirloom Creole of the cowpea genotypes (Sempre Verde, Rabú de tatu, Corujinha, and Paulistinha) and nine cultivars (BRS No-vaera, BRS Olhonegro, BRS Verdejante, BRS Exuberante, BRS Pajeú, BRS Miranda, IPA 206, BRS Tapaihum, and BRS Pingo de Ouro) were processed using four deep learning architectures for feature extraction (vectorization): InceptionV3, SqueezeNet, VGG16, and VGG19. Six machine learning algorithms were evaluated: K-Nearest Neighbors (KNN), Decision Tree, Random Forest (RF), Gradient Boosting (GB), Support Vector Machines (SVM), and Multi-Layer Perceptron (MLP). The MLP (Artificial Neural Network) and SVM models, particularly when integrated with the InceptionV3 embedder, demonstrated superior performance. For pod classification, these models achieved near-perfect performance, with Area Under the Curve (AUC) and Classification Accuracy (CA) of 1.000. For green beans, the MLP maintained high accuracy (CA = 0.977) and better probabilistic calibration (lower Log-Loss) than the SVM. Digital image-based identification associated with machine learning is an efficient, non-destructive approach for the morphological characterization and discrimination of cowpea genotypes, supporting high-throughput phenotyping (HTP) applications.
Why it matches plant phenotyping methodsデジタル画像と機械学習による莢・サヤインゲンの形態的特徴抽出と遺伝子型識別が研究の中心であり、ハイスループット植物表現型解析への応用を明示している。
abstractthe objective was to adjust machine learning models to identify cowpea from digital images of pods and green beans using artificial intelligence techniques.
Apple leaf disease segmentation is critical for yield and quality preservation in what is globally one of the most economically significant fruit crops. Despite recent advances in deep learning, real-world orchard environments present three primary challenges: (1) low contrast between lesions and background textures, which hinders accurate localization; (2) leaf overlap and occlusion, leading to incomplete feature representation and increased false negatives; and (3) the inherent limitations of unimodal RGB imagery in capturing subtle pathological features, which constrains generalization and accuracy. To address these issues, we proposed Language-Infused Visual Mamba (LViM), a dual-path U-Net architecture that integrates Mamba and Transformer modules for semantic-visual feature fusion. LViM achieves robust segmentation in complex environments through three core innovations: (1) A U-shaped Multimodal Transformer (MTT) branch integrated with AMBERT, which leverages inter-modal semantic relationships to enhance textual feature extraction and provide high-level semantic cues, thereby improving lesion-background discriminability; (2) a U-shaped Visual State Space (VMamba) branch that employs 2D Selective Scanning (SS2D) and Visual State Space (VSS) blocks to capture global context and fine-grained details, mitigating the impact of occlusion; and (3) Cross-Attention Gate Fusion (CAGF) and Linguistic Cross-Nested (LCN) modules that facilitate efficient cross-modal alignment and hierarchical feature modeling to better identify subtle lesions. Experimental results demonstrate that LViM consistently outperforms the VM-UNet baseline, yielding improvements of 4.05% in Precision, 4.25% in Dice coefficient, 4.49% in mIoU, and 4.23% in Recall.
Why it matches plant phenotyping methodsリンゴ葉の病斑を画像から分割する手法を開発し、複雑な環境での性能を評価しており、植物病害状態の取得・推定が研究の中心である。
abstractApple leaf disease segmentation is critical for yield and quality preservation
Reproduction assets foundThe paper's curated multimodal apple leaf disease dataset (image-text pairs with pixel-level annotations for four disease types) is explicitly stated as publicly released in the authors' LViM GitHub repository. Code/models are only promised 'upon acceptance,' so the dataset asset qualifies as public, while the code is.Dataset · publicThe curated multimodal apple leaf disease dataset constructed in this study has been publicly released at https://github.com/csuft1906ll/LViMOpen asset ↗csuft1906ll/LViMlines:273-283Plant phenotyping relevance match · UnverifiedEurope PMC · OpenAlex · Crossref · checked 13 Sept 2026
Accurate tree counting from remote sensing data is essential for forest inventory, biomass estimation, carbon accounting, and ecological monitoring. However, existing approaches predominantly rely on airborne RGB imagery and often struggle in complex forest scenes where neighboring crowns exhibit highly similar textures and colors and where overlapping crown boundaries become ambiguous. To address this limitation, the LiDAR-derived Canopy Height Model (CHM) is introduced as a complementary modality that provides explicit cues on canopy height variation and vertical structure to support RGB-based analysis. Building on this, we propose BCAR-Net, a broker-guided RGB and depth (RGB-D) multimodal framework that couples bidirectional cross-modal interaction, adaptive tri-branch fusion, and auxiliary reconstruction within a two-stage optimization scheme. Specifically, a bidirectional cross-attention U-Net generates an intermediate broker RGB-D representation from paired RGB images and depth maps through symmetric bidirectional cross-attention between the two modalities and direction-aware gating. The original RGB image, depth map, and broker representation are then jointly encoded by three weight-sharing branches and adaptively aggregated by a spatial fusion gate for density-map regression. To regularize the fused latent feature, a multi-scale cross-attention reconstruction decoder provides auxiliary RGB and depth reconstruction supervision by querying multi-scale BCA-UNet encoder features through 2D cross-attention, and a reconstruction-oriented first stage replaces externally generated fused-image supervision, yielding a task-consistent optimization scheme. Experiments on the NEONTreeEvaluation benchmark show that BCAR-Net consistently outperforms single-modality settings and direct RGB-D concatenation multimodal baseline. Additional experiments on a public UAV RGB-LiDAR dataset provide a small-scale supplementary evaluation under a different acquisition setting, where BCAR-Net achieves modest but consistent improvements over RGB-only and depth-only baselines. These results demonstrate that the proposed framework offers an effective but computationally cautious solution for tree counting in complex forest environments.
Why it matches plant phenotyping methodsRGB画像とLiDAR由来データから樹木数を推定する深層学習手法を開発し、複数ベンチマークで比較評価しており、植物個体の計測手法が研究の中心である。
abstractwe propose BCAR-Net, a broker-guided RGB and depth (RGB-D) multimodal framework
Abstract Waterlogging is a major constraint on barley productivity, yet its dynamic, multi-phase nature makes it challenging to dissect using traditional phenotyping approaches. High-throughput phenotyping (HTP) platforms address this by enabling temporal, multi-sensor imaging of large populations, but generate complex datasets that demand new analytical frameworks. Here, we imaged 230 barley accessions over 14 days of waterlogging stress and seven days of recovery using visible, chlorophyll fluorescence, and hyperspectral sensors. Explainable AI was applied to classify stress responses into early stress, late stress, and recovery phases, achieving 86% classification accuracy, and to identify the hyperspectral indices most informative for each phase. Water index (WATER1) and structure insensitive pigment index (SIPI) emerged as primary predictors of stress response. Longitudinal genome-wide association studies (GWAS), using a treatment-by-marker interaction model, identified 236 significant loci across 12 linkage disequilibrium blocks, implicating candidate genes involved in oxidative stress regulation, transcriptional control, and auxin transport. MYB transcription factors were consistently identified across all stress phases, underscoring their central role in waterlogging adaptation. To support interpretation of longitudinal GWAS results, we developed 3D-QTLVis, an interactive visualisation tool that extends Manhattan plots across time, enabling clearer identification of dynamic genomic regions underlying stress tolerance.
Why it matches plant phenotyping methods長期マルチセンサー画像による水ストレス応答の表現型取得と、AIによるフェーズ分類・指標抽出が研究の中心であり、3D-QTLVisも開発している。
abstractHigh-throughput phenotyping (HTP) platforms address this by enabling temporal, multi-sensor imaging of large populations
Reproduction assets foundThe paper's authors publicly release their GWAS Interaction model R scripts and the 3D-QTLVis Shiny visualization tool on GitHub; no public phenotype dataset or trained model deposit is stated (phenotypic data only as summary statistics in supplements).Code · publicCode used for running the GWAS interaction model in R and the 3D-QTLVis tool are available at https://github.com/Walshj73/3D-QTLVis .Open asset ↗Walshj73/3D-QTLVislines:216-267Code / dataset availability confirmedEurope PMC · checked 5 Sept 2026
ToBRFV is a major threat to tomato and pepper crops because it spreads quickly and survives for a long time in the environment. Since there are few ways to control it after infection, early detection before symptoms are visible is crucial. Yet, only limited public datasets are available for this research. We present one of the first openly accessible, longitudinal multispectral image dataset dedicated to ToBRFV detection. In this study, two tomato cultivars and two pepper cultivars, all of which are commercially important and widely cultivated in greenhouses, were selected. Using these plants ensures that the dataset reflects real-world agricultural practices and captures variability across commercially grown types. Both healthy and ToBRFV-inoculated plants from each cultivar were included in the imaging process. All plants were cultivated under fully controlled greenhouse conditions in Adana Province, Türkiye. Healthy and infected tomato plants were grown in two separate greenhouses to prevent cross-contamination. Imaging was conducted over a 29-day period using Red-Green-Blue (RGB) and Visible Near Infrared (VNIR) cameras, including narrowband captures at 800 nm and 1000 nm, from multiple viewing angles. Infection status was confirmed via Reverse Transcription quantitative Polymerase Chain Reaction (RT-qPCR) analysis at multiple time points. The dataset is organized into four clean, labelled subsets and released under a CC BY 4.0 license. This resource provides unique opportunities for developing and benchmarking computer vision and machine learning approaches for pre-symptomatic plant disease detection, spectral feature analysis, and integration into precision agriculture systems. By combining controlled experimental design, spectral diversity, and open access, it establishes a robust foundation for cross-disciplinary research in plant pathology, agricultural engineering, and artificial intelligence.
Why it matches plant phenotyping methods植物病害状態を対象にした縦断マルチスペクトル画像データセットであり、公開データセットとして開発・ベンチマーク利用を目的とするため、表現型取得が中心です。
abstractWe present one of the first openly accessible, longitudinal multispectral image dataset dedicated to ToBRFV detection.
Reproduction assets foundThe article is a Data in Brief describing the authors' own openly released longitudinal multispectral plant image dataset (ToBRFV-LMID) for tomato and pepper disease detection, deposited on Zenodo under CC BY 4.0 with a direct DOI URL. This is a paper-specific, public, directly actionable phenotype/image asset. No codeDataset · publicData accessibility
Repository name: ZENODO
Data identification number: 10.5281/zenodo.17244968
Direct URL to data: https://doi.org/10.5281/zenodo.17244968Open asset ↗ZENODO · 10.5281/zenodo.17244968html-lines:98-126Plant phenotyping relevance match · UnverifiedCrossref · checked 5 Sept 2026
Unmanned aerial vehicle (UAV) based remote sensing has emerged as a disruptive technology for detecting crop water stress (CWS) in real time, precisely and at low cost offering significant advancements over conventional approaches. The study examined the red green blue (RGB), multispectral (MSP), hyperspectral (HSP), thermal image sensors integrated with UAVs, which offers a high-spatial and temporal resolution of physiological indicators such as chlorophyll content and canopy cover, canopy temperature, stomatal conductance. The study highlights that in spring maize, random forest (RF) models using UAV-derived MSP and thermal indices with leaf area index (LAI) performed well (R² > 0.575, root mean square error (RMSE)
Why it matches plant phenotyping methodsUAV搭載センサーによる作物の水ストレスや生理形質のモニタリング技術をレビューしており、表現型取得法が中心である。
titleRecent trends in crop water stress monitoring using remote sensing technologies: A review
Abstract Background This study evaluates the performance of two UAV (Unmanned Aerial Vehicle) platforms for vineyard monitoring, focusing on geometric accuracy, effective resolution, and vegetation index consistency. Methods Two UAV (Unmanned Aerial Vehicle) platforms were compared over a vineyard using multiple flights with a consumer-grade drone and a single RTK (Real-Time Kinematic) enabled flight at similar altitude over a single vineyard block under comparable acquisition conditions, followed by photogrammetric reconstruction, DSM (Digital Surface Model) co-registration, effective resolution analysis, and RGB (Red-Green-Blue) based vegetation index comparison. Results This study demonstrates that the integration of RTK (Real-Time Kinematic) technology in the Mavic 3E improves the absolute accuracy of DSMs (Digital Surface Models), eliminating systematic vertical offsets of ~ 35 m observed in the Mavic 2E products. After applying a robust Z-shift (vertical) correction, DSMs (Digital Surface Models) from both platforms became directly comparable, with RMSE (Root Mean Square Error) values reduced to ~ 1.3 m while NMAD (Normalized Median Absolute Deviation) remained stable. RTK (Real-Time Kinematic) positioning in the Mavic 3E ensured reliable absolute georeferencing, whereas DSMs (Digital Surface Models) derived from the Mavic 2 Enterprise Zoom remained internally consistent after post-processing, though with lower absolute positional accuracy. Effective resolution analysis further showed that the Mavic 3E imagery preserves higher spatial detail than the Mavic 2E, underscoring the importance of sensor optics and stability for vineyard monitoring. Comparisons of vegetation indices revealed that normalized indices such as NGRDI (Normalized Green Red Difference Index) and VARI (Visible Atmospherically Resistant Index) provide consistent results across platforms, while ExG (Excess Green Index) exhibited strong biases and wide limits of agreement, reflecting its sensitivity to radiometric differences. Conclusions For cross-platform or multi-temporal monitoring, NGRDI (Normalized Green Red Difference Index) and VARI (Visible Atmospherically Resistant Index) proved to be more robust, whereas ExG (Excess Green Index) should only be applied after radiometric harmonization. This study provides a replicable workflow for UAV (Unmanned Aerial Vehicle) based vineyard monitoring that integrates geometric alignment, DSM (Digital Surface Model) correction, effective resolution assessment, and index comparison, offering practical recommendations for researchers and practitioners aiming to ensure reliable and comparable UAV (Unmanned Aerial Vehicle) derived vineyard metrics.
Why it matches plant phenotyping methodsUAV画像・写真測量・DSM補正・植生指数を比較検証し、ブドウ園の植物状態を再現可能に測定するワークフローを中心に扱っているため、植物フェノタイピング手法研究に該当する。
abstractThis study evaluates the performance of two UAV (Unmanned Aerial Vehicle) platforms for vineyard monitoring, focusing on geometric accuracy, effective resolution, and vegetation index consistency.
Abstract— Plant diseases significantly affect agricultural productivity, food quality, and global food security [1], [16] Traditional disease diagnosis methods rely heavily on human expertise and manual inspection, making the process time-consuming, expensive, and prone to errors [1], [20]. Recent advancements in artificial intelligence and deep learning have transformed plant disease detection systems by enabling automatic, accurate, and real-time identification of plant diseases using digital images [2], [3], [20]. This review paper presents a comprehensive analysis of recent deep learning approaches used for automated plant disease detection. The study discusses various convolutional neural network architectures, transfer learning methods, attention mechanisms, Vision Transformers, and hybrid deep learning techniques applied in agricultural disease diagnosis [2]– [5], Publicly available datasets, evaluation metrics, preprocessing techniques, and comparative analyses of existing methods are also presented. Furthermore, the paper identifies current research challenges such as dataset imbalance, environmental variability, computational complexity, and limited real-world adaptability[16]. Finally, emerging trends including explainable artificial intelligence, federated learning, lightweight edge computing models, drone-based monitoring systems, and multimodal agricultural intelligence are explored [6]. This review aims to provide researchers and practitioners with a detailed understanding of the current state-of-the-art deep learning techniques for intelligent plant disease detection and future research opportunities. Keywords: Plant Disease Detection, Deep Learning, Convolutional Neural Network, Transfer Learning, Precision Agriculture, Computer Vision, Smart Farming, Vision Transformer.
Why it matches plant phenotyping methods植物病害を画像から自動検出する深層学習手法を体系的に比較・レビューしており、植物状態の画像ベース表現型計測が中心である。
abstractThis review paper presents a comprehensive analysis of recent deep learning approaches used for automated plant disease detection.
Accurate identification of tomato lateral shoots is essential for automated pruning and plant monitoring in greenhouse production. However, complex illumination, leaf occlusion, and morphological variability often reduce detection reliability in optical vision systems. This study proposes an optical vision-based framework that integrates deep learning perception with large language model assisted pruning decision support. A tomato lateral Shoot image dataset was constructed using RGB imaging in greenhouse environments. A lightweight YOLOv8n instance segmentation model with the Convolutional Block Attention Module (CBAM) was developed to enhance feature representation. Data augmentation strategies were applied to simulate illumination variations and improve model robustness. Model interpretability was analyzed using Principal Component Analysis (PCA) and Gradient weighted Class Activation Mapping (Grad CAM). Experimental results show that the proposed YOLOv8n-seg+CBAM model achieves a mAP 0.5 of 98.1% with only 3.28M parameters and an average inference time of 8.0 ms per image. Monte Carlo Dropout was further introduced to estimate the spatial uncertainty of cutting points. These structured perception features were provided to a large language model (LLM), enabling context aware pruning decision assistance. The proposed framework integrates vision-based shoot detection, uncertainty estimation, and LLM-assisted reasoning into a unified pipeline, enabling more reliable pruning decisions and improving safety and robustness compared with vision-only approaches in greenhouse environments.
Why it matches plant phenotyping methodsトマト側枝をRGB画像から検出・セグメンテーションし、不確実性推定まで行う画像ベースの植物形態計測手法を開発しており、方法論が中心である。
abstractThis study proposes an optical vision-based framework that integrates deep learning perception with large language model assisted pruning decision support.
The high-precision instance segmentation of tree saplings is a fundamental prerequisite for the high-throughput phenotypic analysis of individual seedlings in intelligent tree breeding and precision silviculture. However, sapling segmentation remains challenging because of blurred boundaries, object adhesion, missed detections, and inaccurate mask delineation in field environments. To improve sapling segmentation performance and address these challenges, this study proposes a multimodal Mask R-CNN framework in which RGB imagery was paired with one multispectral-derived vegetation index at a time to construct separate RGB-VI input combinations, taking ginkgo saplings as a representative case. A dataset of 400 saplings was constructed using a high-throughput field phenotyping platform. The backbone network was extended with an independent vegetation index branch, and three fusion strategies (early, multi-step, and late fusion) were designed within a feature pyramid network to enable multi-scale multimodal feature integration. The results showed that all multimodal models outperformed unimodal baselines in terms of segmentation accuracy and recall. Among them, the multi-step fusion strategy achieved the best performance, while the RGB-EVI multi-step fusion model achieved the highest strict-matching precision (AP@75 = 87.7%) and recall (71.3%), with superior performance in dense sapling delineation and background suppression. These findings indicate that multimodal feature fusion can effectively improve sapling instance segmentation and provide methodological support for high-throughput plant phenotyping.
Why it matches plant phenotyping methodsマルチモーダル画像による樹木苗個体のインスタンスセグメンテーション手法を開発・比較し、高スループット表現型解析を支援することが中心である。
abstractThe high-precision instance segmentation of tree saplings is a fundamental prerequisite for the high-throughput phenotypic analysis of individual seedlings in intelligent tree breeding and precision silviculture.
AppleStrawberryGreenhouseLiDAR / point cloudRGB / grayscaleFruitSegmentationTracking
Accurate and consistent fruit monitoring over time is a key step towards automated agricultural production systems. However, this task is inherently difficult due to variations in fruit size, shape, occlusion, orientation, and the dynamic nature of orchards where fruits may appear or disappear between observations. In this article, we propose a novel method for fruit instance segmentation and re-identification on 3D terrestrial point clouds collected over time. Our approach directly operates on dense colored point clouds, capturing fine-grained 3D spatial detail. We segment individual fruits using a learning-based instance segmentation method applied directly to the point cloud. For each segmented fruit, we extract a compact and discriminative descriptor using a 3D sparse convolutional neural network. To track fruits across different times, we introduce an attention-based matching network that associates fruits with their counterparts from previous sessions. Matching is performed using a probabilistic assignment scheme, selecting the most likely associations across time. We evaluate our approach on real-world datasets of strawberries and apples, demonstrating that it outperforms existing methods in both instance segmentation and temporal re-identification, enabling robust and precise fruit monitoring across complex and dynamic orchard environments. • We propose a new performant approach to autonomous fruit tracking in real greenhouses. • It segments fruits using learning-based instance segmentation and RGB 3D point clouds. • Segmented fruits are encoded by a 3D CNN and matched via attentive data association. • Experiments on real strawberry and apple datasets show our method outperforms others. • Our approach enables precise temporal fruit monitoring in real and complex scenarios.
Why it matches plant phenotyping methods果実を個体単位で3D点群からセグメンテーションし、時系列追跡する画像解析手法の開発・評価が研究の中心であり、植物器官の状態を抽出するため適格。
abstractwe propose a novel method for fruit instance segmentation and re-identification on 3D terrestrial point clouds collected over time
Reproduction assets foundThe paper explicitly states that the authors' implementation of the fruit matching method (IRIS3D) is publicly available on GitHub, which is the computational analysis code for this paper's fruit segmentation and re-identification phenotyping pipeline.Code · publicThe implementation of our fruit matching method is publicly available at https://github.com/PRBonn/IRIS3D .Open asset ↗PRBonn/IRIS3Dlines:72-99Code / dataset availability confirmedCrossref · OpenAlex · Europe PMC · checked 5 Sept 2026
Ear density ( ) and ear surface area in cereals are important traits for adaptation to low inputs and climate change. Here we propose a high-throughput field phenotyping method to estimate these traits using nadir and 45° RGB images acquired by the Phenomobile ground robot. First, the YOLOv5 ear detection algorithm is applied to nadir RGB images to estimate . Second, an ear segmentation algorithm is applied to nadir and 45° RGB images to compute the ear gap fraction at different viewing angles. The Beer-Lambert law is then inverted to compute the ear area index (EAI) from the observed ear gap fraction. is finally derived as the ratio between EAI and . We applied the methodology to a panel of 10 commercial bread wheat varieties how both traits vary across 12 environments. The relative error obtained for awnless varieties is 12% (56 ears m -2 ) for and 18% (1.3 cm 2 ) for . For awned varieties, ground-truth observations of were shown to be biased due to an overestimation of awns contribution, leading to an error of 41% (3.6 cm 2 ). was strongly correlated with grain dry mass per ear at harvest ( r 2 = 0.80 across genotypes and environments, r 2 per genotype ranged between 0.80 and 0.95) and was strongly correlated with grain yield ( r 2 = 0.83). These results indicate that both EAI and can be interesting non-destructive proxies for yield and grain dry mass per ear.
Why it matches plant phenotyping methodsRGB画像と地上ロボット、物体検出・セグメンテーション・Beer–Lambert法を組み合わせ、コムギ穂の密度と表面積を推定・検証する手法が研究の中心であるため。
abstractHere we propose a high-throughput field phenotyping method to estimate these traits using nadir and 45° RGB images acquired by the Phenomobile ground robot.
Reproduction assets foundThe authors publicly release their ear surface area estimation algorithm with an example dataset on an INRAE forge repository, and the Phenomobile-derived ear density/ear surface area estimations used in the multi-environment analysis are included as supplemental material with the open-access article. The YOLOv5 GWC_SoDataset · publicThe algorithm developed to estimate the EAI and the average ear surface using binary images from ear segmentation are publicly available in the repository https://forge.inrae.fr/raul.lopez-lozano/wheat-ear-surface , jointly with an example dataset from the Mauguio 2023 trial (4 treatments, 1 replicate). The Phenomobile estimations of ear surface area and ear density used in the multi-environmental mixed model presented in Section 2.5 are included as supplemental material.Open asset ↗lines:614-652Code / dataset availability confirmedCrossref · Europe PMC · checked 5 Sept 2026
Three-dimensional (3D) point-cloud phenotyping enables non-destructive and repeatable characterization of plant architecture, supporting the measurement of traits such as internode length, branching topology, and organ orientation. This article presents TomatoPGT (Tomato Plant Graph Twin) , a 3D tomato dataset designed for research on semantic/instance segmentation, graph-based structural representation, and graph-derived phenotypic trait extraction. The dataset contains 42 scans from three greenhouse-grown tomato plants acquired across early to mid-vegetative development using a rotational multi-view imaging system. Each scan consists of 60-70 overlapping RGB images captured under uniform illumination and reconstructed into a metrically scaled dense colored point cloud using Structure-from-Motion and multi-view stereo. TomatoPGT provides: (i) multi-view RGB images, (ii) dense colored point clouds, (iii) manually curated semantic and instance annotations at organ level, (iv) graph representations encoding plant topology and geometry, and (v) tabulated phenotypic traits computed deterministically from the graphs (internode length, insertion angles, and phyllotactic angles). TomatoPGT supports reproducible development and evaluation of 3D phenotyping pipelines, including learning-based segmentation and graph-based modeling of plant architecture.
Why it matches plant phenotyping methods植物の3D形態表現型抽出を目的としたデータセットで、画像・点群・器官アノテーション・グラフ・形質値を提供し、再現可能なフェノタイピング手法の開発と評価を直接支援している。
abstractThis article presents TomatoPGT (Tomato Plant Graph Twin) , a 3D tomato dataset designed for research on semantic/instance segmentation, graph-based structural representation, and graph-derived phenotypic trait extraction.
Reproduction assets foundThe paper's own TomatoPGT dataset (multi-view RGB images, dense point clouds, semantic/instance annotations, graph representations, and CSV phenotypic traits) is publicly deposited on Mendeley Data, and the authors' Cloud-Seg/Cloud-Graph software tools plus supplementary materials (camera calibrations, example datasetsDataset · publicRepository name 1: Mendeley[2].
Data identification number: DOI: 10.17632/72md54c7n7.1
Direct URL to data: https://data.mendeley.com/datasets/72md54c7n7/1Open asset ↗Mendeley · 10.17632/72md54c7n7.1html-lines:105-178Code · public6. Code and documentation: CloudSeg and CloudGraph software tools, environment specifications, and example usage instructions are hosted on Zenodo[3].Open asset ↗Zenodohtml-lines:264-308Plant phenotyping relevance match · UnverifiedCrossref · OpenAlex · checked 14 Sept 2026
Accurate prediction of nitrogen utilization efficiency (NUtE) is critical for breeding nitrogen-efficient crop cultivars and optimizing field nitrogen management. Traditional prediction methods are time-consuming and limited to resolving plant-level nitrogen dynamics, hindering effective phenotype acquisition at the field scale and understanding of nitrogen uptake and transport in crops. This study aims to couple proximal remote sensing (PRS) and a crop growth model (CGM, i.e., WheatGrow model) via high-throughput phenotyping (HTP) techniques to establish a non-destructive prediction framework for plot-level NUtE at the field scale. Firstly, the organ-level nitrogen submodule was developed and integrated into the WheatGrow model, improving the simulation accuracy of plant nitrogen accumulation dynamics. Second, proximal RGB data combined with a deep-shallow machine learning approach enabled high-precision estimation of organ-specific critical nitrogen concentrations (leaf: R 2 = 0.94, RMSE = 0.21%; spike: R 2 = 0.95, RMSE = 0.10%). Fitted parameters of critical nitrogen dilution curves (CNDCs) demonstrated variations between cultivar and management in both organs, with spike nitrogen dilution rates exhibiting greater sensitivity to management practices than leaves. Finally, coupling PRS-derived organ-specific CNDCs with the enhanced WheatGrow model through the ensemble Kalman filter (EnKF) algorithm, yielded precise NUtE predictions at a small spatial scale (RMSE = 4.54 kg kg −1 , Bias = 0.05). Validation across multi-year, multi-cultivar trials demonstrated robust performance, reducing NUtE prediction errors below 10% (RRMSE = 9.9% ± 0.8%). This framework bridges HTP techniques with crop modeling, may catalyze a paradigm shift in CGMs from empirical parameterization to real-time sensing, and advance scalable nitrogen use efficiency phenotyping in sustainable crop improvement and smart agriculture.
Why it matches plant phenotyping methods高スループット表現型計測、近接リモートセンシング、RGB画像、機械学習、作物モデルを統合し、器官別窒素形質と圃場スケールのNUtEを推定する方法を開発・検証しており、表現型取得が研究の中心である。
abstractThis study aims to couple proximal remote sensing (PRS) and a crop growth model (CGM, i.e., WheatGrow model) via high-throughput phenotyping (HTP) techniques to establish a non-destructive prediction framework for plot-level NUtE at the field scale.
Plant diseases are one of the main limiting factors in global agricultural productivity, causing significant losses and compromising food security. The increasing complexity of production systems and the limitations of traditional diagnostic methods, based mainly on visual assessment and laboratory analyses, have driven the incorporation of artificial intelligence (AI) in plant pathology. In this context, the present study aimed to synthesize the advances, challenges, and gaps related to the application of AI in the detection, monitoring, and forecasting of plant diseases. This is an integrative literature review, conducted through systematic searches in national and international scientific databases, encompassing studies that addressed machine learning techniques, deep learning, and hybrid models applied to plant pathology. Approaches based on RGB images, multispectral and hyperspectral data, integration with unmanned aerial vehicles (UAVs), and forecasting models based on climatic variables were analyzed. The results show that convolutional neural networks and temporal architectures, such as LSTM, have substantially increased the diagnostic accuracy and forecasting potential of the systems, especially when integrated with environmental data. However, challenges persist related to the generalization of the models, scarcity of representative databases, field variability, and high computational cost. It is concluded that AI represents a strategic tool for the transition from a from a reactive phytopathology to a predictive and decision-support approach. However, its consolidation under real cultivation conditions depends on robust agronomic validation, methodological standardization, and multidisciplinary integration, aiming at more precise, sustainable systems applicable to precision agriculture.
Why it matches plant phenotyping methods植物病害の画像・マルチスペクトル・ハイパースペクトル観測とAIによる検出・予測を主題とするレビューであり、植物状態（病害）の取得・推定手法が中心です。
abstractthe present study aimed to synthesize the advances, challenges, and gaps related to the application of AI in the detection, monitoring, and forecasting of plant diseases.
Crop cultivation sustains the livelihoods of a substantial portion of households across developing economies, yet the crops on which those households depend are perpetually at risk from pathogenic infections that erode both yield volume and produce quality. Infected fields, when not addressed at the right time, translate into mounting financial strain that smallholder growers — who operate with limited financial reserves — are ill-equipped to withstand. The dominant method of spotting such infections today still relies on a farmer walking the field and judging leaf condition by eye, or waiting for an agronomist's visit — a workflow that is neither fast nor consistent enough for large-scale cultivation. Progress in deep learning has fundamentally changed what automated visual inspection can accomplish, and plant pathology diagnosis is one of the fields that has benefitted most visibly. Image-based pipelines can now scan a leaf photograph and return a disease classification in fractions of a second. Among the architectures driving this capability, Convolutional Neural Networks occupy a central role: their layered filter design allows them to extract and encode visually informative features — texture discontinuities, color anomalies, lesion geometry — without any manual specification of what to look for. We present a CNN-driven plant disease recognition system that operates on photographs of plant leaves and returns a disease label together with a confidence estimate. Our training corpus is the Plant Village benchmark collection, a large repository of annotated leaf images spanning healthy and diseased specimens across multiple crop varieties. The input pipeline applies spatial normalization, pixel rescaling, and augmentation strategies to condition the data before it reaches the network. A React.js browser interface connects end users to the model via a lightweight prediction API, enabling diagnosis without any specialist involvement. Validation results affirm that this deep learning approach surpasses rule-based image processing baselines on both accuracy and response time, and the system holds clear potential for adoption in early disease management program
Why it matches plant phenotyping methods植物葉画像から病害状態を推定するCNNベースの画像解析手法を開発し、既存手法との性能比較・検証を行っており、表現型取得が研究の中心である。
abstractWe present a CNN-driven plant disease recognition system that operates on photographs of plant leaves and returns a disease label together with a confidence estimate.
While unmanned aircraft system (UAS)-based photogrammetry and light detection and ranging (LiDAR) are increasingly used for canopy height estimation in forestry and other orchard systems, their application to pecan orchards remains limited. Accurate measurements of tree height and canopy structure are essential in pecan production for assessing tree growth and health, and for supporting precision orchard management. This study provides one of the first systematic evaluations of UAS-based structure-from-motion (SfM) photogrammetry and UAS-mounted LiDAR for estimating pecan tree height. A rotary-wing UAS equipped with RGB and near-infrared (NIR) cameras collected imagery at 60 and 120 m aboveground over two pecan orchards containing 480 and 308 trees, and LiDAR data were acquired at 70 m. UAS imagery was processed to generate three-dimensional (3D) point clouds, digital surface models (DSMs), digital terrain models (DTMs), and orthomosaics. DTMs were derived using point cloud classification and DSM filtering, and tree heights were calculated relative to these terrain models using canopy height models (CHMs) and point cloud–based approaches. LiDAR data were processed to produce calibrated point clouds, DSMs, and DTMs, from which tree heights were extracted using comparable methods. Image-based tree heights showed strong agreement with manual measurements, with point cloud–derived high percentiles or maxima [ R 2 = 0.982–0.996; root mean square error (RMSE) = 14 to 25 cm] consistently outperforming CHM-based estimates across ground elevation methods, camera types, and flight altitudes. LiDAR-derived tree heights exhibited similarly high accuracy. Image-based and LiDAR-derived heights were strongly correlated across all trees at 120 m ( R 2 = 0.982–0.995; RMSE = 18–25 cm), confirming the reliability of SfM photogrammetry. However, incomplete canopy reconstruction in some 60 m datasets led to underestimation, highlighting the importance of sufficient image overlap for accurate 3D canopy modeling. These results demonstrate that UAS image-based point clouds can provide pecan tree heights comparable to LiDAR, offering a cost-effective approach for tree growth monitoring, orchard management, and precision agriculture applications.
Why it matches plant phenotyping methodsUAS画像測量とLiDARを用いた pecan 樹高推定法を系統的に比較・検証しており、植物形態形質の取得が研究の中心である。
abstractThis study provides one of the first systematic evaluations of UAS-based structure-from-motion (SfM) photogrammetry and UAS-mounted LiDAR for estimating pecan tree height.
Precise, non-destructive detection of fruit maturity is a cornerstone of modern precision agriculture, directly impacting harvest scheduling and post-harvest quality control. In the case of strawberries (Fragaria × ananassa), in-field automated assessment is persistently hampered by the fruit’s diminutive size, subtle physiological colour transitions, and frequent occlusion by foliage. To overcome these limitations, we developed SMLO-YOLO, a specialised lightweight vision system designed to reliably detect different maturity stages on edge devices under complex orchard conditions. The proposed architecture incorporates a Cross-Scale Aggregation Neck (HDP-Neck) driven by entropy-guided dynamic sampling, which effectively concentrates computational resources on fruit regions while filtering background noise. Additionally, we introduce a Shape-aware Intersection-over-Union (ShapeIoU) loss and a Boundary- and Class-aware Knowledge Distillation (BCKD) strategy to specifically address the challenge of detecting overlapping clusters and low-maturity fruits. Validation on custom datasets collected from commercial orchards in Sichuan and Shanxi demonstrated that the final SMLO-YOLO model, after BCKDloss-based knowledge distillation, achieved an mAP50 of 92.4% at an inference speed of 256.41 FPS, with 6.49 M parameters and 15.0 GFLOPs. These metrics indicate that the system successfully balances high-throughput detection with the non-harvestable low-maturity fruits of agricultural robotics, offering a robust tool for objective, real-time maturity monitoring.
Why it matches plant phenotyping methodsイチゴ果実の成熟段階という植物器官の状態を画像から推定する軽量YOLO手法を開発し、実圃場データで性能検証しており、フェノタイピング手法が研究の中心です。
abstractwe developed SMLO-YOLO, a specialised lightweight vision system designed to reliably detect different maturity stages on edge devices under complex orchard conditions.
Agricultural crop diseases can greatly reduce production quality and overall farm output, making early identification important for sustainable farming. This study introduces a smart agricultural rover that applies a multimodal deep learning approach for real-time crop disease monitoring in field environments. The proposed system gathers RGB images, thermal information, and environmental measurements such as temperature, humidity, and soil moisture through integrated sensors connected to a Raspberry Pi 4. For on-device analysis, a lightweight TensorFlow Lite (TFLite) model is utilized to classify crop diseases efficiently at the edge. To improve detection performance under different illumination conditions, the system evaluates both original and CLAHE-enhanced images using a dualinference mechanism supported by entropy and confidence-based decision metrics. The rover is implemented on a mobile robotic platform equipped with motor control and battery support to enable autonomous movement in agricultural fields. By combining sensor fusion, edge intelligence, and robotic mobility, the developed system supports accurate identification of diseases such as Powdery Mildew and Rust, helping farmers take preventive action and improve crop management practices
Why it matches plant phenotyping methodsRGB・熱画像とセンサ融合、エッジ推論、画像強調による作物病害検出システムを開発しており、植物の病害状態を推定する方法が中心である。
abstractThis study introduces a smart agricultural rover that applies a multimodal deep learning approach for real-time crop disease monitoring in field environments.
Timely harvesting of fresh tomatoes is urgently needed. To address this issue, this study proposes DDC-YOLOv11n, a model suitable for real-time detection of tomato ripeness in complex greenhouse environments. A Zero-DCE adaptive enhancement module is first deployed at the input stage to restore and enhance the true color and texture details of the images. An improved Deep Residual Shrinkage Network (DRSN) is then added to YOLOv11n to perform adaptive soft-threshold filtering on feature maps, reducing the interference of image noise on the detection targets. Finally, the CBAM spatial attention is enhanced through dilated convolution and channel grouping to form the LKCBAM module, which expands the equivalent receptive field while controlling the increase in parameters, thereby improving tomato detection accuracy in occluded and dense scenes. Experimental results show that the DDC-YOLOv11n model achieves the best recognition performance: compared with the original YOLOv11n, its mAP@0.5, precision, recall, and F1 score are increased by 16.8%, 24.6%, 8.3%, and 18.1%, respectively. These findings facilitate real-time tomato ripeness detection in complex greenhouse environments and provide perceptual information for subsequent management tasks such as harvesting.
Why it matches plant phenotyping methodsトマト果実の成熟度という植物器官の状態を画像から推定するモデルを開発・評価しており、フェノタイピング手法が研究の中心である。
titleA real-time ripeness detection model for tomatoes in complex greenhouse environments.
The accurate quantification of glucoraphanin (GRA), a crucial health-promoting compound in broccoli, is vital for assessing its nutritional quality. However, traditional methods relying on destructive laboratory assays hinder rapid quality monitoring. To address this limitation, we developed a novel non-destructive, multimodal deep learning framework that integrates two phenotypic data modalities—image-based phenotypes from red-green-blue (RGB) leaf images and field-measured plant morphological traits—for accurate GRA estimation. Our proposed model, Parallel-Enhanced FasterNet (PE-FasterNet), incorporates two key innovations: a Gated Parallel Routing Attention (GPRA) mechanism for enhanced feature extraction, and a Phenotype-Guided Cross-Attention Feature Fusion (PG-CAFF) module for effective cross-modal fusion. Through rigorous evaluation, the model achieved a standard random-split test R 2 of 0.985 and a Leave-One-Group-Out (LOGO) cross-validation R 2 of 0.979, demonstrating highly accurate and generalized GRA predictions. This performance represents a substantial improvement over state-of-the-art convolutional neural network (CNN) and Vision Transformer models, affirming the architectural superiority of our approach. This study not only provides a robust tool for rapid, non-destructive prediction of GRA but also demonstrates a viable pathway toward data-driven crop quality management and precision breeding in broccoli.
Why it matches plant phenotyping methodsブロッコリー葉画像と形態形質からグルコラファニンを非破壊推定する深層学習法を開発・検証しており、表現型取得・抽出ワークフローが研究の中心である。
abstractwe developed a novel non-destructive, multimodal deep learning framework that integrates two phenotypic data modalities—image-based phenotypes from red-green-blue (RGB) leaf images and field-measured plant morphological traits—for accurate GRA estimation.
The early detection and spatial characterization of crop damage are critical for improving decision-making in precision agriculture, particularly in regions where traditional monitoring methods are limited in scalability and objectivity. This study presents an integrated information processing framework that couples UAV-based image acquisition, instance segmentation, slicing-aided inference of large orthomosaics, and georeferenced spatial analysis into a single reproducible pipeline for the detection and mapping of crop damage. The framework is applied to maize cultivated under traditional milpa systems in Yucatán, Mexico, a region characterized by intercropping, irregular plant spacing, and complex backgrounds rarely represented in mainstream agricultural deep learning benchmarks. High-resolution RGB images were systematically acquired over maize fields in Yucatán, Mexico, and curated into specialized datasets representing parcels, individual plants, and damaged vegetation. Instance segmentation models based on the YOLOv11 architecture were trained and evaluated to extract visual information related to crop condition, while the Slicing-Aided Hyper Inference (SAHI) method was integrated to enable efficient processing of large orthomosaic images. The proposed framework achieved high performance in detecting maize plants, with a precision of 92.9% and an mAP50 of 94.2%, and demonstrated reliable identification of damage patterns associated with Spodoptera frugiperda, reaching a precision of 79.2% and an mAP50 of 71.7%. The resulting georeferenced outputs provide spatially explicit information that supports quantitative analysis of crop health and damage distribution. The results indicate that the proposed framework constitutes a scalable and reproducible approach for UAV-based visual information extraction, with potential applicability to broader agricultural monitoring and data-driven decision support systems.
Why it matches plant phenotyping methodsUAV画像、インスタンスセグメンテーション、SAHIを統合し、作物の損傷状態を抽出・定量化する再現可能な手法が研究の中心であり、性能評価も実施している。
abstractThis study presents an integrated information processing framework that couples UAV-based image acquisition, instance segmentation, slicing-aided inference of large orthomosaics, and georeferenced spatial analysis into a single reproducible pipeline for the detection and mapping of crop damage.
A comprehensive description of plant phenotypes of certain taxa is an important task when describing genera and species, as well as when setting their natural taxonomies. The development of modern technologies of effective phenotyping makes it possible to obtain a large amount of data with a quantitative and/or qualitative description of various traits in plants, mainly based on the analysis of their digital images. The study compared the results of the F2 hybrids assessment - visually and using machine learning methods - of two endemic tetraploid (2n = 4x = 28) wheat species which are Ethiopian wheat (Triticum aethiopicum Jakubz.) and Kartalian or Dika wheat (T. carthlicum Nevski). In the latter case, it is proposed to use the method of a mixture of Gaussian (normal) distributions in plant morphometry in order to identify groups that differ in character values. Most taxonomically important (species-specific) traits are controlled oligogenically and have a clear phenotypic manifestation, so hybridological analysis was an indispensable and basic type of analysis for subsequent detailed phenotyping of wheat spikes using machine-learning methods. According to a number of criteria, the estimates of patterns of inheritance obtained by different methods coincide. Based on the conducted research, we can state that the trait "tetraaristatum" (the presence of awns on both flower and spike glumes) is species-specific (taxonomically important) for T. carthlicum and it can be effectively used for taxonomic purposes both in carrying out hybridological analysis and in experiments using machine learning. Such a species-specific character is the "character (type) of awnedness" for T. aethiopicum. Our study demonstrates that a combination of automatic phenotyping methods and a model of a mixture of Gaussian distributions can, in principle, lead to an automatic analysis of the allocation of classes in F2 hybrids. It allows, in turn, to detect the presence of genes associated with species-specific traits of wheat plants. Further, the improvement of the applied artificial intelligence (AI) algorithms is required.
Why it matches plant phenotyping methodsコムギ穂の形態形質を対象に、画像に基づく機械学習フェノタイピングとガウス混合モデルを提案・適用しており、表現型の自動抽出・分類が研究の中心である。
abstractThe study compared the results of the F2 hybrids assessment - visually and using machine learning methods
Reproduction assets foundThe paper's supplementary materials (Supplementary Tables S1–S3 and Figure S1) contain the paper-specific phenotyping data: species-specific trait descriptions, the 19 spike morphometric characters per projection, and the Gaussian mixture model splitting results (means, variances, group sizes, χ² values). The full textSupplement · publicof these traits
are controlled by oligogenes and have a clear phenotypic
manifestation, the hybridological method was an indispensable
and primary type of analysis for subsequent detailed
phenotyping spikes of wheat species using machine learning
methods.
Supplementary Materials are available in the online version of the paper:
https://vavilov.elpub.ru/jour/manager/files/Suppl_Kruch_Engl_30_3.pdf
Plant material. The object of study was interspecific
hybrids obtained by crossing two endemic tetraploid
wheat species ♀T. aethiopicum Jakubz. (k-19301/2) with
♂T. carthlicum
Nevski (k-32496). The experiment was produced
in spring sowing in the greenhouses of the Breeding
and Genetics Complex (BGC)Open asset ↗lines:111-200Code / dataset availability confirmedEurope PMC · checked 5 Sept 2026
PollenBB16 is an RGB pollen image dataset of Chilean flora with pixel-accurate instance segmentation masks, whose annotation was fully verified by an expert palynologist to guarantee the taxonomic reliability of every published instance. The dataset is designed to close a concrete gap in existing palynological datasets, which typically combine low taxonomic diversity, few samples per class, and low-resolution crops restricted to bounding boxes. PollenBB16 contains 16,198 brightfield optical microscopy images at the native resolution of 3088 × 2064 pixels and 36,383 pixel-accurate polygons across 16 species from the Biobío Region, spanning endemic, native and exotic species of high ecological and melliferous value such as Eucryphia glutinosa and Quillaja saponaria (endemic), Gevuina avellana and Aristotelia chilensis (native), and Medicago sativa and Brassica rapa (introduced). Each spatial position is recorded at three focal planes. The displacement along the z axis reveals features of the exine together with information on the internal structure of the grain that remain inaccessible on a single plane. From this multifocal information, more robust convolutional networks can be trained with more accurate classification. The operational quality of the dataset is backed by a leakage-safe partition that keeps the three focal planes of the same position in the same subset to avoid metric inflation, complemented by a YOLO11n-seg baseline trained for 50 epochs that reaches 0.985 mask mAP@50 on the validation set, establishing a reproducible reference point. Beyond deep learning, PollenBB16 enables interdisciplinary applications in aerobiology, biodiversity monitoring under climate change, ecological restoration of the South American temperate forest, and botanical-origin authentication of Chilean monofloral honeys.
Why it matches plant phenotyping methods植物由来の花粉粒を対象とした高解像度画像データセットとセグメンテーション基準を構築し、深層学習による画像解析を再現可能な形で検証しているため、植物フェノタイピング手法・データセットとして中心的です。
abstractPollenBB16 is an RGB pollen image dataset of Chilean flora with pixel-accurate instance segmentation masks
Reproduction assets foundThe paper's PollenBB16 pollen image dataset (16,198 multifocal RGB microscopy images with pixel-accurate instance segmentation masks) and the accompanying authors' script polygons_to_bboxes.py are publicly deposited on Zenodo (10.5281/zenodo.19830051), per the Data Availability Statement.Code · publicThe only custom code distributed with this Data Descriptor is the Python script polygons_to_bboxes.py, which regenerates the YOLO bounding-box labels in labels_bb/ from the polygon labels in labels/. The script is packaged inside the scripts/ folder of the same Zenodo repository that hosts the dataset (10.5281/zenodo.19830051) and is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license, with no restrictions on access.Open asset ↗Zenodo · 10.5281/zenodo.19830051html-lines:731-797Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
The rapid and nondestructive classification of maize kernels is of great significance for seed screening and quality evaluation. Existing hyperspectral image classification methods based on the Mamba architecture can effectively represent spectral and spatial features; however, they still face limitations in time-frequency analysis and multimodal feature fusion. In addition, traditional approaches often rely heavily on spectral preprocessing, which may introduce additional errors and compromise the model's robustness and generalization ability. To address these challenges, this paper proposes a novel cross-modal classification framework named CD-TriMamba, which jointly leverages hyperspectral data and visible-light images for comprehensive feature extraction and deep fusion. Specifically, an innovative feature extraction module is designed, consisting of a Spectral Curvelet Convolution (SCC) module for hyperspectral data and a Curvelet-Decomposed Convolution (CDC) module for spatial modeling. A feature rearrangement mechanism is further introduced to mine critical information from both spectral and spatial modalities. Finally, a ConvNeXt-guided tri-branch cross-fusion structure (TriMamba) is constructed to achieve deep collaboration and efficient integration between spectral and spatial features. Experimental results demonstrate that the proposed model achieves outstanding performance in seed classification, with an accuracy (Acc) of 99.2% and a Kappa value of 99.1%. These results strongly confirm the effectiveness and broad application potential of cross-modal feature fusion in maize kernel classification.
Why it matches plant phenotyping methodsマルチモーダル画像からトウモロコシ種子の健全性を推定する新規分類フレームワークを開発しており、種子状態の取得・抽出手法が研究の中心である。
abstractthis paper proposes a novel cross-modal classification framework named CD-TriMamba, which jointly leverages hyperspectral data and visible-light images for comprehensive feature extraction and deep fusion.
The increasing global food insecurity driven by climate-induced natural hazards and soil degradation has made the resilience of alternative agricultural systems a critical focus in risk management. This study presents a geospatially integrated monitoring framework, the Optimized Multi-Scale Adaptive Graph Neural Network (OMSA-GNN), designed to mitigate risks associated with nutrient instability in hydroponic and aeroponic environments. The proposed system leverages a Raspberry Pi-based IoT network to monitor complex interactions among microclimatic variables, plant physiological health, and nutrient concentrations, treating them as localized geospatial data points. To enhance decision-making under environmental uncertainty, an Improved Sparrow Search Algorithm (ISSA) is employed to optimize the predictive performance of the GNN. The OMSA-GNN model incorporates visual plant indices as a proximal remote sensing approach to enable early detection of physiological stress that may lead to crop failure. Evaluated using a lettuce growth dataset, the framework demonstrates superior performance in forecasting growth trajectories and managing resource-related risks compared to conventional static models. The results highlight a scalable approach for improving the reliability of urban food systems, where traditional land-based agriculture is increasingly vulnerable to natural hazards.
Why it matches plant phenotyping methods植物の生理的ストレスと成長軌跡を、視覚的植物指数およびIoTセンサーデータから推定するGNNベースの監視・解析手法が研究の中心であり、植物表現型取得と予測に該当する。
abstractThe OMSA-GNN model incorporates visual plant indices as a proximal remote sensing approach to enable early detection of physiological stress that may lead to crop failure.
Reproduction assets foundThe paper's Data Availability statement points to a public Kaggle lettuce growth dataset used for evaluation, matching an allowed URL. No author code or model checkpoints are disclosed.Dataset · publicThe datasets used and/or analyzed during the current study are available in the Kaggle repository, https://www.kaggle.com/datasets/jurijsruko/lettuce/data.Open asset ↗Kaggle · jurijsruko/lettucehtml-lines:469-500Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · checked 15 Sept 2026
Abstract Plant breeding is essential for crop improvement, yet progress is often hindered by slow, laborious, and subjective field phenotyping methods. High‐throughput phenotyping (HTP), particularly image‐based methodologies powered by machine learning, offers a pathway to overcome these limitations. However, achieving robustness and generalization when analyzing diverse genotypes within a crop and across reproductive stages remains challenging and can affect model performance and the accurate extraction of phenotypic features. This study evaluated the performance of semantic segmentation models across a diverse panel of genotypes and distinct crop reproductive stages, using wheat ( Triticum aestivum L.), sorghum ( Sorghum bicolor L.), and corn ( Zea mays L.) as case studies. The primary objectives were to analyze (i) the overall prediction performance on the aggregated dataset for each crop, (ii) the stratified performance by genotype and collection date, and (iii) the temporal and genotypic transferability across growth stages and unseen genotypes. Four distinct smartphone cameras were used to collect images of the reproductive structure across crop growth stages (different collection dates) from 160 corn, 80 sorghum, and 40 wheat genotypes. The total number of images per crop was 2000 for wheat, 4000 for sorghum, and 3840 for corn. Five semantic segmentation models were tested in this study—DeepLabv3+, MaskFormer, SegFormer, SegNet, and U‐Net—using the images and respective binary masks for training and testing. The SegFormer model achieved the highest intersection over union (IoU) values for corn (0.90) and sorghum (0.92), while the U‐Net model performed best for wheat (0.89). A minor performance decline, with IoU differences up to 0.1, was observed when testing the same model across different genotypes. However, the temporal transferability drops up to 0.5 IoU when training and inferring on different crop growth stages. The main reason for those changes may lie in the natural color and organ architecture temporal changes between the trained and tested datasets when transferring the models across growth stages. These results highlight the urgent need to prioritize robustness and transferability when developing reliable in‐field HTP methodologies.
Why it matches plant phenotyping methods植物の生殖器官画像からの表現型抽出に用いるセマンティックセグメンテーション手法を、作物・遺伝子型・生育段階間で性能と転移性の観点から比較検証しており、方法論が研究の中心である。
abstractThis study evaluated the performance of semantic segmentation models across a diverse panel of genotypes and distinct crop reproductive stages
Understanding diurnal canopy orientation in crops is important for interpreting plant responses to light and environmental conditions, yet field-based quantification remains limited. In this study, we present Heliocot, a field RGB imaging approach that converts time-resolved images into reference-area standardized projected leaf area (PLA) time series to quantify within-day canopy orientation dynamics in early-season cotton. Leaf instance segmentation was performed using YOLOv8m-seg and refined through a 144-combination post-processing optimization. On the held-out early-stage validation/tuning set, the selected workflow showed strong agreement with manual ground truth (R2 = 0.948; NRMSE = 0.082) and destructive leaf area measurements (R2 = 0.836). Derived diurnal metrics, including Daily Orientation Amplitude (DOA) and Peak Orientation Index (POI), consistently revealed a midday maximum (13:15) in canopy projection. Exploratory genotype-level analysis suggested negative associations between orientation indices and selected plant traits, including specific leaf area (SLA) versus DOA (r = −0.71, p = 0.021, R2 = 0.508), destructive leaf area (LA) versus DOA (r = −0.69, p = 0.028, R2 = 0.471), and stem dry weight (SDW) versus POI (r = −0.74, p = 0.014, R2 = 0.554), while plant height was not significantly associated with POI and DOA (p > 0.05). Although currently limited to early-season conditions and two field-imaging dates, this approach provides a practical workflow for field-based monitoring of canopy projection dynamics in cotton, while broader temporal and environmental validation remains necessary.
Why it matches plant phenotyping methods圃場RGB画像から葉面積と日周キャノピー配向動態を推定する手法を開発・検証しており、植物形質取得が研究の中心である。
abstractwe present Heliocot, a field RGB imaging approach that converts time-resolved images into reference-area standardized projected leaf area (PLA) time series to quantify within-day canopy orientation dynamics in early-season cotton.
This article presents the development and experimental study of a compact, automated apple sorting machine based on computer vision, indirect fruit weight estimation, and color grading. It is designed for use in small and medium-sized farms, where the use of industrial lines is limited by high cost and complex maintenance. The proposed machine enables non-destructive assessing and sorting of apples in a continuous flow mode without the use of mechanical weighing, includes a fruit feeding and positioning module, a computer vision system, an image processing unit, and a sorting actuator synchronized with the conveyor movement. Fruit weight and color assessment is based on visual geometric parameters (diameter, fruit height, projected area, and the proportion of surface color), extracted from digital images, followed by classification by product category using regression models. As part of the experimental study, a correlation analysis was conducted between the actual weight of apples and their geometric parameters for five varieties typical of Kazakhstan. It was shown that the projected fruit area exhibits the most stable correlation with weight, justifying its use as the primary predictor in constructing a regression model for indirect weight estimation. An assessment of the accuracy of apple classification by product categories was carried out and the influence of conveyor speed on the stability and correctness of sorting was observed. The experimental results with a total 1250 apples of five varieties - Aport Alexander, Sinap Almaty, Kazakhski Yubileinyi, Ainur, and Nursat indicated that the optimal operating mode for the machine is an apple transport speed of 0.16 m/s. In this mode, sorting throughput is approximately 400 kg/hour, with an average accuracy of 92% for the automatic classification in accordance with GOST requirements. These results confirm that the proposed approach provides sufficient real-time sorting accuracy with a simple machine design. The machine can be used as a standalone sorting solution, as well as a base platform for further expansion of functionality by integrating surface defect assessment and grade identification modules.
Why it matches plant phenotyping methods果実の形態・色・重量を画像から推定し分類するコンピュータビジョン方式と装置の開発が中心であり、単なる品質測定ではなく、再利用可能な植物器官形質の取得・推定法を提示している。
abstractThis article presents the development and experimental study of a compact, automated apple sorting machine based on computer vision, indirect fruit weight estimation, and color grading.
The research proposes Cross Disease Similarity Awareness Learning (CDSAL), a robust multiclass tomato leaf disease detection framework based on high-quality and explainable deep learning. The approach solves the problem of superimposed patterns of disease especially Leaf Miner, Tomato Spotted Wilt Virus (TSWV), and nutrient deficiencies through the combination of multi-domain feature learning and inter-disease similarity modeling. In contrast to conventional metric learning or contrastive learning methods that function on pairwise or triplet sample associations, CDSAL develops a class-level Cross Disease Similarity Matrix that represents structured inter-disease proximity within the embedding space. Moreover, rather than employing episodic prototype construction typical of few-shot learning, the proposed system persistently updates centroid representations throughout supervised training and incorporates similarity-aware regularization directly into the loss function. This facilitates structural embedding reshaping specifically designed for visually overlapping illness categories, beyond traditional prototype-based learning methodologies. The input images are processed through HSV based green masking, morphological cleaning, extraction of leaf contours and resizing, and using a large amount of geometric and color-space augmentation to reduce the imbalance among the classes. DenseNet121 and EfficientNet-B0 are used to obtain feature representations and class-separated centroid of latent embedding's to form a Cross Disease Similarity Matrix, where similarity-aware optimization is possible during training. Grad-CAM on the target layers offers decipherable disease-specific activation signatures. The findings of the experiments show that classification accuracy at unseen samples is 99.77% with high resilience to visual confounding. The predictions, proximity of diseases that are similar and explainable features are provided by CDSAL, thereby facilitating reliable decision-making in agricultural diagnostics.
Why it matches plant phenotyping methodsトマト葉の病害状態を画像から推定する深層学習手法を提案し、前処理・特徴抽出・類似度学習・説明可能性を技術的中心として評価しているため。
abstractThe research proposes Cross Disease Similarity Awareness Learning (CDSAL), a robust multiclass tomato leaf disease detection framework based on high-quality and explainable deep learning.
Reproduction assets foundThe paper's plant-phenotyping inputs are two publicly available Kaggle image datasets explicitly named in the Data Availability statement: PlantVillage (emmarex/plantdisease) used as the main dataset and TomatoVillage (mamtag/tomato-village) used for ablation/field-condition experiments. No author analysis code, modelsDataset · publicThe datasets analyzed during the current study are available in the Kaggle repository. [https://www.kaggle.com/datasets/emmarex/plantdisease]Open asset ↗Kaggle · emmarex/plantdiseasehtml-lines:605-624Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
The rapid and accurate quantification of plant phosphorus (P) content is essential for the real-time assessment of crop P status and improvement of P fertilizer use efficiency. However, non-destructive and rapid approaches for P monitoring are limited. In this study, the feasibility of monitoring plant phosphorus content (PPC) in winter wheat was evaluated through multi-source feature fusion of unmanned aerial vehicle (UAV) imagery based on a long-term field experiment with five P treatments. Multiple spectral features, including color indices (CIs), fractional vegetation cover (FVC), vegetation indices (VIs), texture features (TFs) and texture indices (TIs), were extracted from UAV RGB and multispectral images. Sensitive spectral features were systematically screened using Pearson correlation analysis, random forest (RF) importance ranking, and the Relief algorithm. Selected features were then fed into three machine learning models, RF, support vector machine (SVM), and k-nearest neighbor (KNN) to predict PPC. The results showed that GRI, VARI, MGRVI, TGI, NDRE, and CIred edge were highly correlated with PPC at the maturity stage (r = 0.96). Both TFs and TIs demonstrated stronger correlations with PPC at the 750 and 840 nm bands, with most TIs outperforming TFs, confirming the feasibility of spectral-based PPC estimation. Based on the selected input variables including DTI (450-Ent, 750-Mea), 840-Mea, and RVI, the SVM model achieved the best performance (R 2 c=0.94, RMSEc=0.29, RPDc=4.03; R 2 v=0.92, RMSEv=0.36, RPDv=3.48). These results highlight the potential of combining VIs, TFs, and TIs features for training machine learning models for PPC prediction, while the organ-level physiological explanations warrantee further investigations under controlled P gradients. This study provides data-driven insights for UAV-based monitoring of plant P nutritional status under local experimental conditions.
Why it matches plant phenotyping methodsUAV画像から特徴量を抽出し、機械学習で冬コムギの植物リン含量という生理形質を推定する手法の開発・評価が中心であり、方法論的検証も実施している。
abstractMultiple spectral features, including color indices (CIs), fractional vegetation cover (FVC), vegetation indices (VIs), texture features (TFs) and texture indices (TIs), were extracted from UAV RGB and multispectral images.
Abstract Plant diseases are a serious danger to the world’s food security, because they lower agricultural output and increase economic losses. Due to subjectivity, fluctuating lighting, and environmental unpredictability, traditional visual examination techniques are frequently incorrect. The Excess Green (ExG) vegetation index and pseudo-thermal representations produced from RGB pictures are two synthetically developed complementary representations that are integrated with RGB imagery in this study’s lightweight multimodal deep learning system to address these issues. Histogram shifting and pseudo-infrared color mapping are used in a reproducible picture alteration pipeline to create the pseudo-thermal modality, which allows for extra visual signals without the need for specific thermal sensors. In order to classify plant diseases while preserving computational efficiency, the suggested framework uses MobileNetV3-Small backbones to extract modality-specific characteristics. This is followed by feature-level fusion. The publicly accessible Ginger Leaf Dataset, which includes RGB pictures of ginger leaves in four different conditions—Damage-Pest, Dehydrated, Healthy, and Leaf-blight—was used for the experiments. For training, validation, and testing, the dataset was split using a stratified 70:15:15 split. Python-based preprocessing procedures were used to create the extra modalities (ExG and pseudo-thermal representations) from the original RGB images. The experimental results show that the combination of the representations with RGB images can enhance the classification performance compared with the unimodal RGB-based models. Ablation experiments are also conducted to examine the contributions of different modalities to the overall categorization accuracy. The experimental results show that plant disease recognition can be improved with the help of efficient computing by combining lightweight convolutional neural networks with computationally generated visual representations.
Why it matches plant phenotyping methodsRGB画像からExG・疑似熱画像を生成し、植物葉の病害状態を分類するマルチモーダル手法が研究の中心であり、アブレーション評価も実施している。
titleHybrid deep learning-based multimodal framework for plant leaf disease classification using RGB, Excess Green (ExG), and pseudo-thermal representations with MobileNetV2
Reproduction assets foundThe paper's phenotyping experiments use the publicly available Ginger Leaf Dataset (RGB leaf images of four ginger leaf conditions), with a public GitHub repository and dataset website. The authors' derived ExG/pseudo-thermal representations and preprocessing scripts are only available upon request, so they do not yetDataset · publicor multispectral images
IEEE Geosci. Remote Sens. Lett. 2025
10.1109/LGRS.2025.XXXXXXX
Ulku, I., Tanriover, O. O. & Akagündüz, E. Cross-band correlation-aware interactive fusion for multispectral images. IEEE Geosci. Remote Sens. Lett.
10.1109/LGRS.2025.XXXXXXX
(2025).
10. Wong, J. Ginger Leaf Dataset. GitHub Repository (2023). https://github.com/wongjay1941/Ginger-Leaf-Dataset
11.
Bhakta I
A novel plant disease prediction model based on thermal images using modified deep convolutional neural network
Precis. Agric. 2023 24 23 39
10.1007/s11119-022-09927-x
Bhakta, I. et al. A novel plant disease prediction model based on thermal images using modified deep convolutional neural network. Precis.Open asset ↗https://github.com/wongjay1941/Ginger-Leaf-Datasetlines:580-681Plant phenotyping relevance match · UnverifiedCrossref · checked 15 Sept 2026
Published20 May 2026Journal of Advanced Computational Intelligence and Intelligent InformaticsCited by 0 · OpenAlex ↗
The real-time quantitative estimation of herbaceous plant growth status holds significant potential for investigating fertilization effects, predicting growth curves, and enhancing crop yield. This study constructed a growth quantification model using an improved YOLOv5 architecture integrated with 3D point cloud processing, with pak choi as an exemplar crop. To improve the recognition accuracy while reducing the number of parameters, we employed a lightweight YOLOv5 model enhanced with Atrous Spatial Pyramid Pooling and Ghost convolution modules for individual pak choi plant localization and growth stage classification. We also developed a segmentation method based on the HSV color space to segment leaves. To estimate the total fresh weight of individual plants, we first calculated the leaf surface area by generating a triangular mesh from the corresponding leaf point clouds and predicted the chlorophyll content using a stacking ensemble model. Subsequently, to address the leaf occlusion issues, the leaf pixel ratio in the images, leaf surface area, and mean leaf chlorophyll content were collectively used as independent variables. Finally, a multiple linear regression model was developed to accurately estimate the total fresh weight of individual pak choi plants. Experimental results demonstrate that the modified YOLOv5 architecture achieves a 3.5% improvement in mAP@0.5 (reaching 96%) and a 4.66% increase in F1-score (attaining 90.26%), while significantly reducing the computational complexity compared to the baseline model. Statistical tests verified that the fitted equation could explain 79% of the variation in the total fresh weight, with an average relative error of 12.16%. This enables non-contact and accurate measurement of the pak choi growth status.
Why it matches plant phenotyping methodsYOLOv5、3D点群、葉面積・クロロフィル推定を統合し、個体の生体重という植物形質を非接触推定する手法が研究の中心である。
abstractThe real-time quantitative estimation of herbaceous plant growth status holds significant potential
Introduction Colorimetric analysis of food using the CIELab/Ch colour space (i.e., from digital images of samples) is an accessible, non-destructive method for carotenoid and anthocyanin content prediction. Literature presents very well-fit, but rudimentary, models for pigment estimation (e.g., single/multiple linear regressions). However, standardised methods that statistically account for the high multicollinearity between CIELab/Ch colour parameters, varying light conditions and colour calibration, and samples with high genotypic variability are lacking. Methods An image analysis optimisation was developed for the prediction of carotenoid and anthocyanin content of 16 carrot genotypes of different colours. Samples were photographed under six light conditions with a digital camera and image colour was calibrated before analysis with the CIELab/Ch colour space. Total pigment contents and individual carotenoid contents were analysed chemically via spectrophotometry and high-performance liquid chromatography, respectively. Partial least squares (PLS) regressions were used to assess the colour-pigment relationships to correct for high multicollinearity amongst the independent variables (CIELab/Ch colour parameters). Results/discussion The PLS models achieved satisfactory accuracy for the prediction of total carotenoid content ( ca. R 2 = 0.77) and total anthocyanin content ( ca. R 2 = 0.81) under all light conditions. The two models are suggested as robust approaches to total pigment prediction with multi-dimensional colour spaces, varying light conditions, and for a sample group of high genotypic variability. The carrot samples proved to have very high genetic diversity within each cultivar, resulting in unsatisfactory models for prediction of individual carotenoids ( ca. R 2 = 0.45) under the default light condition. However, all the results can be used to expand databases (towards artificial intelligence) and aid breeding programmes in search for higher concentrations of these interesting antioxidants for human health.
Why it matches plant phenotyping methodsニンジン試料の画像色解析を最適化し、化学分析値を用いてカロテノイド・アントシアニン含量を予測する手法を開発・検証しており、植物形質取得が研究の中心である。
abstractThe PLS models achieved satisfactory accuracy for the prediction of total carotenoid content ( ca. R 2 = 0.77) and total anthocyanin content ( ca. R 2 = 0.81) under all light conditions.
Reproduction assets foundThe authors deposited the paper's data and protocols in public repositories (DOI links in the Data availability statement). The anthocyanin quantification protocol is explicitly linked (10.34894/BTPTSV), and the other two DOIs (10.34894/P37WCL, 10.34894/OUURRH) are stated to hold the paper's data. No separate author's'Dataset · publicData and protocols are available in the following links: https://doi.org/10.34894/P37WCL , https://doi.org/10.34894/OUURRH , https://doi.org/10.34894/BTPTSV .Open asset ↗10.34894/P37WCLlines:641-686Dataset · publicData and protocols are available in the following links: https://doi.org/10.34894/P37WCL , https://doi.org/10.34894/OUURRH , https://doi.org/10.34894/BTPTSV .Open asset ↗10.34894/OUURRHlines:641-686Plant phenotyping relevance match · UnverifiedOpenAlex · checked 15 Sept 2026
GrapevineNeRF / 3D Gaussian SplattingLiDAR / point cloudRGB / grayscaleFruitStem / branchPose / keypoint estimation2D/3D reconstructionSegmentation
• End-to-end pipeline from neural reconstruction to physical grape berry manipulation. • Efficient point clouds generation using NeRF and the metric scale derived directly from robot kinematics. • RANSAC sphere fitting achieves 92.1% berry detection precision without annotated training data. • Stem-aligned 6-DoF pose optimization improves end-to-end grip success by 17.2%. Table grape thinning requires selective removal of 20–40% of berries from dense clusters. In practice, workers decide which berries to remove by considering both the approximate berry count and local 3D spatial characteristics such as crowding and relative positioning. Automating this task is challenging because conventional 2D image-based approaches suffer from occlusion-related counting errors and lack explicit 3D spatial information necessary for reliable manipulation. We propose a robot-integrated vision pipeline that reconstructs grape bunch structure from posed multi-view RGB images. Neural Radiance Fields (NeRF) is used to learn a volumetric scene representation, from which a dense, low-noise point cloud is extracted via depth back-projection, and RANSAC-based geometric fitting models individual berries and stems, enabling berry-level segmentation and orientation estimation for manipulation planning. The perception pipeline uses an eye-in-hand RealSense D405 camera mounted on a Fanuc CRX-5iA collaborative robot. Camera poses are derived from the robot kinematic chain, allowing the reconstructed point cloud and detected berry centers to be expressed directly in the metric robot base frame without external scale recovery. In robot-mounted RealSense D405 experiments on 10 grape bunches, RANSAC sphere fitting achieved a counting MAE of 0.50 berries, RMSE of 0.71 berries, and mean center localization error of 2.71 mm. On a 52-bunch benchmark, RANSAC sphere fitting outperforms the learning-based SoftGroup++ method for 3D berry instance segmentation (92.1% vs 82.8% average precision) without requiring annotated training data. In 35 manipulation trials, the system achieved an 85.7% pre-grasp reachability rate and an 83.3% conditional target success rate demonstrating an end-to-end pipeline from neural reconstruction to manipulation-ready berry poses.
Why it matches plant phenotyping methodsブドウ房の3D再構成、ベリー分割・計数・位置推定を中核とするロボット統合型フェノタイピング手法であり、技術性能も定量評価している。
abstractWe propose a robot-integrated vision pipeline that reconstructs grape bunch structure from posed multi-view RGB images.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · Crossref · checked 5 Sept 2026
Abstract Unmanned aerial vehicle (UAV)-based phenotyping has been applied to assess potato traits, however, its use to identify canopy traits associated with tuber yield across diverse genotypes remains limited. The objective of this study was to evaluate the use of UAV-based field phenotyping, integrating RGB and thermal imaging, to identify key canopy traits associated with tuber yield and it´s agronomic components in a set of eight potato genotypes grown across two environments and two growing seasons. Despite higher seasonal rainfall in Chiloé, tuber yields were consistently greater in Osorno, underscoring that total precipitation alone is less important than its temporal distribution and effective crop water availability; this makes it necessary to supplement with irrigation during the period of highest demand. RGB-derived vegetation indices and canopy temperature successfully differentiated genotypes, although their discriminatory power varied according to developmental stage and environmental conditions, with intermediate to late growth stages generally providing the strongest genotype separation. Canopy temperature supplied complementary physiological information related to canopy water status, whereas RGB traits captured broader variation in canopy structure and greenness. These findings highlight the importance of integrating phenological stage and environmental context when interpreting remote sensing data, and demonstrate the strong potential of UAV-based HTP to support breeding and agronomic strategies aimed at improving drought resilience, yield stability, and selection efficiency in potato.
Why it matches plant phenotyping methodsUAVによるRGB・熱画像を用いた圃場フェノタイピングが中心で、ジャガイモのキャノピー形質を抽出・評価し、遺伝子型間比較や収量関連性を検討している。
abstractevaluate the use of UAV-based field phenotyping, integrating RGB and thermal imaging, to identify key canopy traits associated with tuber yield
Introduction: Wheat kernel hardness, vitreousness, and creaseness are key determinants of milling performance, yet they reflect different physical scales of grain structure and are not necessarily coupled. Methods: We developed a digital phenotyping framework based on hyperspectral imaging and spectral unmixing to quantify these traits at both kernel and cultivar levels in a diverse panel of common wheat. Pixel-level spectral unmixing resolved glassy, intermediate, and mealy endosperm components within individual kernels, enabling vitreousness to be expressed as a continuous spatial index. Results: The hyperspectral-derived vitreousness index showed moderate associations with kernel protein content and the protein-to-starch ratio, consistent with variation in endosperm packing density, but weak relationships with kernel hardness and crease geometry. Kernel hardness, primarily determined by puroindoline genotype, showed limited association with bulk protein and starch composition. Crease geometry, quantified using composite indices from RGB images, captured macroscopic grain features largely independent of both hardness and vitreousness. Discussion: These results demonstrate that hardness, vitreousness, and creaseness represent complementary but largely independent dimensions of grain quality, corresponding to molecular-scale adhesion, mesoscale packing, and macroscopic geometry, respectively. The proposed framework provides a scalable, non-destructive approach for resolving intra-kernel heterogeneity, enabling improved digital phenotyping for wheat breeding and quality assessment.
Why it matches plant phenotyping methodsハイパースペクトル画像とスペクトルアンミキシングを用いて小麦粒の硝子質を定量するデジタル表現型解析フレームワークを開発しており、形質取得手法が中心的である。
abstractWe developed a digital phenotyping framework based on hyperspectral imaging and spectral unmixing to quantify these traits at both kernel and cultivar levels in a diverse panel of common wheat.
Reproduction assets foundThe paper's data availability statement deposits full hyperspectral image cubes and RGB image datasets on Figshare, and the supplementary material includes Python analysis scripts (Supplementary Code S1–S2) and processed feature tables (Supplementary Table S3) directly reproducing the paper's phenotyping measurements.Dataset · publicfull hyperspectral image cubes and associated RGB imagedatasets are available via Research Datas 1 – 3 at Figshare: https://doi.org/10.6084/m9.figshare.31259530Open asset ↗Figshare · 10.6084/m9.figshare.31259530lines:151-201Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
To achieve high-precision and high-efficiency estimation of the Leaf Area Index (LAI) of jujube trees using drone remote sensing, and to overcome the limitations of traditional vegetation index methods, such as saturation in the later stages of crop growth, sensitivity to background noise, and difficulty in capturing temporal dynamics, this study proposes a parallel hybrid deep learning framework using CNN-GRU. The model adaptively extracts spatial-spectral local features from drone RGB images through a convolutional neural network (CNN) branch, while a gated recurrent unit (GRU) branch learns the sequential evolution of LAI during key phenological periods. Finally, a meta-learner integrates spatial-temporal information for decision-making. To verify the model's effectiveness and prediction performance, the study systematically collected multi-temporal ground-measured LAI data and synchronized drone remote sensing images during critical growth stages of jujube trees in two independent years: 2024 (Bachu, Xinjiang) and 2025 (Alaer, Xinjiang). A series of spectral and texture indices were extracted as model inputs. The experimental results show that the proposed CNN-GRU model exhibits excellent learning and fitting capabilities on the training set, with an R 2 value of 0.839. On the test set, after optimization with data augmentation strategies, the model's prediction accuracy is significantly improved, with prediction accuracy reaching its best level, with an R 2 of 0.83 and an RMSE of 0.150. All error metrics outperform mainstream comparative models such as Transformer, KNN, MLP, and CNN. This study demonstrates that the hybrid deep learning architecture, combining spatial feature extraction and time-series modeling, is an effective approach for accurate and robust remote sensing inversion of crop LAI in complex agricultural scenarios, providing a reliable technical tool for the digital management of smart orchards and precise agricultural decision-making.
Why it matches plant phenotyping methodsUAV RGB画像からナツメ樹のLAIを推定するCNN-GRU手法を開発し、複数年データと比較モデルで性能検証しており、植物形質取得が中心である。
abstractthis study proposes a parallel hybrid deep learning framework using CNN-GRU
AppleField / plotRGB / grayscaleFruitClassificationObject detectionSegmentationGrowth / development / phenology
This article presents a comprehensive dataset of 1406 RGB images of apples ( Malus domestica ), covering three key growth stages-immature (green), semi-mature (color transition), and mature (red). The dataset serves as a resource for detecting and segmenting apples across different developmental phases. Each image includes pixel-level instance segmentation masks annotated in JSON format using the VGG Image Annotator (VIA), ensuring compatibility with deep learning frameworks. The dataset's real-world variability-spanning lighting conditions, occlusions, and clustered fruit arrangements-enhances its utility for training generalizable computer vision models in precision agriculture. It supports tasks such as fruit detection, segmentation and growth-stage classification, addressing the scarcity of annotated data for transitional maturity phases. With 2574 annotated apple instances, this dataset facilitates research on maturity grading and transfer learning for agricultural robotics. By standardizing annotations and incorporating diverse field conditions, this dataset reduces preprocessing overhead and accelerates the development of deployable AI solutions for orchard management. It is particularly valuable for improving model robustness in heterogeneous environments, thereby advancing data-driven horticultural practices.
Why it matches plant phenotyping methodsリンゴ果実の発育段階・成熟度という植物器官の状態を対象に、画素単位アノテーション付き画像データセットを構築しており、観測・抽出手法の再利用可能な基盤が中心である。
abstractThis article presents a comprehensive dataset of 1406 RGB images of apples ( Malus domestica ), covering three key growth stages-immature (green), semi-mature (color transition), and mature (red).
Reproduction assets foundThe paper is a data descriptor for a public apple image dataset (1406 RGB images, pixel-level instance segmentation masks in JSON) deposited on Mendeley Data with a direct URL and DOI, matching the allowed URL exactly.Dataset · publicRepository name: Wang, Dandan; Wang, Bo (2026), “A Multi-Stage, Pixel-Level Annotated Apple Dataset for Precision Agriculture Research”, Mendeley Data, V4
Data identification number: 10.17632/gfcmdbvw65.4
Direct URL to data:https://data.mendeley.com/datasets/gfcmdbvw65/4Open asset ↗Mendeley Data · 10.17632/gfcmdbvw65.4html-lines:1-97Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · checked 5 Sept 2026
Abstract Background: The stink bug complex is one of the most damaging pests of soybean, reducing yield and seed quality. Genetic resistance remains the most sustainable and effective management strategy, but its quantitative inheritance and labor-intensive field phenotyping make its implementation in breeding programs challenging. Objective: This study explored high-throughput phenotyping (HTP) using unmanned aerial vehicles (UAVs) equipped with RGB cameras to evaluate a soybean population and the potential of phenotyping to stink bug resistance by correlating image-derived features and machine learning (ML) models. Methods: A population of 304 soybean lines was evaluated in alpha-lattice design trials across two seasons under natural infestations. Five resistance-related traits, grain yield (GY), healthy seed weight (HSW), number of days to maturity (NDM), tolerance (TOL), and leaf retention (LR), were manually scored and linked to UAV-derived vegetation indices (VIs) and texture indices (TIs). Three ML models (AdaBoost, SVM, MLP) were tested to predict these traits from aerial features. Results: Results showed that VIs, particularly Visible Atmospherically Resistant Index at the 25th percentile (VARI_P25), were consistently associated with resistance-related traits, while decision tree analysis highlighted TIs at 45° and 135° as complementary sources of structural information. Prediction ability was highest for GY, HSW, and NDM, especially in flights near flowering and maturity, but remained low for TOL and LR. Integrating multiple flights modestly improved accuracy, whereas cross-season predictions were unreliable. Nonetheless, indices such as VARI_P25 provided useful cross-season correlations for HSW and TOL, enabling early screening of less promising lines. Conclusion: This pioneering study demonstrates that UAV–ML pipelines can capture genetic signals of stink bug resistance in soybean, despite environmental complexity. These findings open new avenues for resistance phenotyping, supporting more efficient breeding strategies and accelerating genetic gains in soybean improvement.
Why it matches plant phenotyping methodsUAV画像と機械学習を用いて、ダイズの抵抗性関連形質や収量を推定するHTPパイプラインを技術的に評価しており、表現型取得・推定法が研究の中心である。
abstractThis study explored high-throughput phenotyping (HTP) using unmanned aerial vehicles (UAVs) equipped with RGB cameras to evaluate a soybean population and the potential of phenotyping to stink bug resistance by correlating image-derived features and machine learning (ML) models.
Accurate and reproducible assessment of foliar disease severity is essential for evaluating the performance of heterogeneous plant communities and understanding host-pathogen interactions. However, traditional visual scoring methods remain subjective, with limited precision, and difficult to scale in large phenotyping experiments. Here, we present a semi-automated image analysis workflow designed to quantify multiple foliar disease symptoms simultaneously on wheat flag leaves sampled from varietal mixtures. The workflow combines three methodological components: (i) a standardized protocol for leaf sampling and imaging, (ii) supervised machine learning segmentation using Random Forest implemented in Ilastik to classify multiple symptoms (powdery mildew and yellow rust), and (iii) a graphical user interface facilitating pipeline deployment by non-specialist operators. To evaluate the influence of image representation on classification performance, four color spaces (RGB, HSV, HLS, LAB) were systematically compared. The approach was validated using images of durum wheat flag leaves collected from a field experiment assessing eight-way varietal mixtures under natural fungal pressure. Cross-validation against manually annotated images demonstrated high segmentation accuracy across all symptom. Comparison among color spaces revealed only minor differences in performance. Overall, this workflow offers a cost-effective, annotation-efficient and reproducible alternative to deep learning approaches, leveraging open-source and actively maintained tools while requiring limited training data and enabling objective, reproducible and scalable disease phenotyping.
Why it matches plant phenotyping methods葉の病害症状を画像解析で定量化するワークフローを開発し、色空間比較と手動アノテーションによる検証を行っており、植物表現型取得法が中心である。
abstractwe present a semi-automated image analysis workflow designed to quantify multiple foliar disease symptoms simultaneously
Reproduction assets foundThe paper's authors explicitly state that all code implementing the leaf disease quantification workflow (SegLeaf, including the graphical interface and documentation) is hosted in a public GitHub repository. No separate public phenotype dataset or trained model checkpoint is described in the supplied blocks.Code · publicted by the Agence Nationale de
la Recherche (ANR) (project SCOOP, grant no. ANR-19-CE32-0011; and project MOBIDIV,
grant no. ANR-20-PCPA-0006).
Code and Data Availability
The method and associated scripts developed in this work are freely available to the re-
search community. All code is hosted in a public GitHub repository at https://github.com/titouanlegourrierec/SegLeaf, which includes the full implementation of the method includ-
ing the graphical interface and documentation to guide users through the analysis pipeline.
15
.
CC-BY 4.0 International license
made available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display tOpen asset ↗titouanlegourrierec/SegLeafpdf-raw-page:15 lines:1-39Code / dataset availability confirmedCrossref · Europe PMC · checked 5 Sept 2026
Abstract Timely and accurate plant disease detection is important for enhancing agricultural productivity and promoting sustainability. The study introduces Multimodal Adaptive Fuzzy-based Deep Neural Network (MAF-DNN) for classification of plant diseases. The proposed method combines fuzzy logic with multimodal data fusion to effectively address the complex interactions and uncertainties in agricultural datasets. The MAF-DNN employs a robust adaptive fuzzy framework with dynamic rule optimization and integrates Hyperspectral Imaging Data (HID) with RGB imaging data to acquire detailed spectral information and high-resolution visual cues for disease classification. The multimodal fusion enhances the model’s ability to capture intricate patterns that relate to plant health, improving the accuracy of disease classification. The experimental results showed that the MAF-DNN outperforms traditional models by achieving an accuracy of 97.8%, precision of 96.5%, recall of 98.2%, and F1-score of 97.3%. Additionally, the adaptive design reduces computational overhead, increases efficiency, and improves scalability for large-scale agricultural applications. The MAF-DNN represents a significant advancement in plant disease classification and provides a robust and efficient solution for precision agriculture.
Why it matches plant phenotyping methods植物病徴を画像から分類するマルチモーダル画像・深層学習手法の開発と性能評価が中心であり、植物の病害状態を直接推定するため。
abstractThe study introduces Multimodal Adaptive Fuzzy-based Deep Neural Network (MAF-DNN) for classification of plant diseases.
Reproduction assets foundThe paper uses two public Kaggle plant disease image datasets (New Plant Diseases Dataset and CCMT Plant Disease Dataset) as its phenotyping inputs and states that the authors' custom MAF-DNN code is publicly available on GitHub, with all three URLs given in the article and matching allowed URLs.Code · publicThe custom code used to develop and evaluate the proposed Multimodal Adaptive Fuzzy Deep Neural Network (MAF-DNN) framework is publicly available at: https://github.com/skbsangeetha/MAF-DNN-Plant-disease-classificationOpen asset ↗skbsangeetha/MAF-DNN-Plant-disease-classificationhtml-lines:102-118Plant phenotyping relevance match · UnverifiedCrossref · checked 15 Sept 2026
Bananas are among the most widely cultivated fruits worldwide and constitute a major staple crop in many developing regions. Banana production is highly vulnerable to foliar diseases, which significantly reduce yield and economic sustainability. Early and accurate disease diagnosis plays a central role in mitigating crop losses. This study presents an image-based artificial intelligence framework for banana leaf disease diagnosis using advanced segmentation and deep learning techniques. Gaussian Kernel Fuzzy C-Means (GKFCM) clustering is employed for pixel-level segmentation, enabling precise isolation of diseased regions by preserving nonlinear boundary characteristics and minimizing intra-cluster variance. For disease classification, a Correlation Weight Bayesian SqueezeNet (CWBSN) model is introduced, integrating Bayesian optimization and correlation-aware feature weighting to enhance discriminative representation and classification robustness. The framework is evaluated using RGB images obtained from the BananaLSD dataset and field-acquired imagery collected at Bangabandhu Sheikh MujiburRahman Agricultural University (BSMRAU), Bangladesh. Performance is assessed using accuracy, precision, recall, F1-score, Matthews Correlation Coefficient (MCC), and Receiver Operating Characteristic (ROC) analysis. Experimental results demonstrate that the proposed GKFCM–CWBSN framework achieves reliable and consistent disease recognition based solely on visual symptom analysis. The study establishes a computational foundation for image-driven plant disease diagnosis, with potential adaptability to future multimodal agricultural sensing systems.
Why it matches plant phenotyping methodsバナナ葉の病斑領域を画像から抽出・分類する手法の開発が研究の中心であり、植物病害状態の表現型推定に該当する。
abstractThis study presents an image-based artificial intelligence framework for banana leaf disease diagnosis using advanced segmentation and deep learning techniques.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 5 Sept 2026
Nitrogen (N) management is critical for optimizing growth and fruit quality in open-field strawberry cultivation, demanding advanced technological solutions for reliable nutrient assessment. However, visual symptom diagnosis, though widely utilized for nutrient monitoring, is inherently subjective and prone to observer bias, resulting in inconsistent and often unreliable assessments. While available accurate tissue analysis is destructive and costly. Nondestructive, in-field imaging techniques such as the normalized difference vegetation index (NDVI) exist but require expensive multispectral imaging systems. To address these limitations, this study developed a streamlined methodology for in-field N status detection using deep learning on standard RGB images. The experiment utilized 'Yotsuboshi' strawberries in a randomized complete block design with sufficient nitrogen (T1) and deficient nitrogen (T2) treatments. To mitigate ambient light variability, a key challenge in open-field phenotyping, a low-cost phenotyping cylinder was developed for standardized smartphone image acquisition. Rigorous four-stage annotation criteria were also introduced to classify the nitrogen status in strawberry leaves as NormalN, LowN, or AdvancedLowN, ensuring a high-quality novel dataset. A YOLO11 model trained on this dataset achieved precision, recall, and mAP50 values exceeding 99%. Subsequent testing using the phenotyping cylinder yielded a mAP50 of 87%. In-field validation without a phenotyping cylinder also demonstrated robust performance under diffuse cloudy conditions (82.7% mAP50), outperforming direct sunlight (79% mAP50). Moreover, the model's classifications of 'NormalN' and 'LowN' statuses strongly corresponded with NDVI measurements, validating the accuracy of the RGB-based approach. This research demonstrates the significant potential of combining deep learning and phenotyping cylinder to create a rapid, low-cost, nondestructive and reliable tool for in-field nitrogen detection, with possible application across different crops and environmental conditions.
Why it matches plant phenotyping methods植物の窒素状態をRGB画像から推定する深層学習法、標準化撮像用デバイス、アノテーション基準、データセットを開発し、NDVIおよび圃場条件で検証しており、表現型取得・推定手法が中心である。
abstractthis study developed a streamlined methodology for in-field N status detection using deep learning on standard RGB images.
Potato crop is highly vulnerable to abiotic stresses like salinity and low nutrient availability. Rapid identification of stress-resilient genotypes is therefore essential for breeding, yet conventional phenotyping is often slow, space-demanding and expensive. We present LOCOPOTS — a LOw-COst high-throughput screening platform for in vitro POTatoes under abiotic Stress — which combines individual in vitro plant culture, low-cost RGB imaging and machine-learning-based automatic segmentation using a trained model of a convolutional neural network, based on U-Net architecture. LOCOPOTS enabled the automated extraction of growth, colour, and vegetation-index traits and demonstrated robust performance across independent phenotyping rounds. We screened 30 potato varieties under control, low-nutrient and saltinity conditions, identifying contrasting growth and physiological responses. Integrated traits such as final area and height, Area_AUC and height_AUC, together with GLI, Ch ol , cive and chlorophyll fluorescence parameters, discriminated genotype performance under stress. Metabolic profiling further revealed genotype-specific reprogramming in carbon and nitrogen metabolism under low nutrition and salt stress, including changes in fructose, myo-inositol, β-aminobutyric acid, γ-aminobutyric acid, proline, and certain polyamines, identifying them as specific chemical biomarkers of plant stress responses. LOCOPOTS provides a scalable, affordable and space-efficient platform for early screening of potato genetic diversity and identification of candidate traits associated with stress resilience.
Why it matches plant phenotyping methods低コストRGB撮像とU-Netによる自動セグメンテーションを中核とする、ジャガイモ表現型取得プラットフォームの開発・検証であり、形態・色・植生指数形質を自動抽出している。
abstractWe present LOCOPOTS — a LOw-COst high-throughput screening platform for in vitro POTatoes under abiotic Stress — which combines individual in vitro plant culture, low-cost RGB imaging and machine-learning-based automatic segmentation using a trained model of a convolutional neural network, based on U-Net architecture.
Unmanned aerial vehicles (UAVs) are broadly used for high-throughput plant phenotyping, yet their long-term use in public-sector research is increasingly challenged by regulatory restrictions and reliance on proprietary platforms. This study presented a regulation-compliant, modular multi-sensor unmanned aerial system (UAS) designed to deliver flexible, high-quality phenotyping data without dependence on restricted ecosystems. A dual-mount, open-architecture payload integrated RGB, multispectral, and thermal sensors, enabling simultaneous acquisition of structural, spectral, and thermal information within a unified workflow. Field validation in a lantana (Lantana camara) breeding trial demonstrated high-precision multi-sensor data fusion and reliable trait extraction. Spatial co-registration achieved centimeter-level accuracy, with alignment errors of 0.88 cm (multispectral) and 3.23 cm (thermal) relative to the RGB reference. UAV-derived canopy height closely matched ground measurements (R2 up to 0.98; RMSE as low as 1.57 cm), while canopy coverage estimates showed consistency across sensing modalities (R2 = 0.99; RMSE = 0.02 m2). Calibrated thermal orthomosaics provided robust canopy temperature estimation (RMSE = 3.13 °C), supporting a quantitative assessment of plant physiological status. Together, these results demonstrate that a regulation-compliant, open-architecture UAV platform can achieve high accuracy in multi-modal phenotyping while maintaining flexibility and cost efficiency. This work demonstrates a scalable and sustainable framework for UAV-based phenotyping, enabling researchers to adapt to evolving regulations while advancing data-driven crop improvement.
Why it matches plant phenotyping methods植物フェノタイピング用のマルチセンサーUAVプラットフォームを設計・検証し、植物形質の抽出精度を評価しているため、方法が中心的である。
abstractThis study presented a regulation-compliant, modular multi-sensor unmanned aerial system (UAS) designed to deliver flexible, high-quality phenotyping data without dependence on restricted ecosystems.
Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · Europe PMC · checked 5 Sept 2026
Efficient phenotyping monitoring of cauliflower is crucial for its breeding and production. However, traditional manual measurement methods are time-consuming and labor-intensive, and existing deep learning (DL) methods mostly focus on the seedling stage, lacking systematic research covering the entire growth period. In this study, RGB images of cauliflower from seedling to harvest were collected. Through systematic screening and evaluation of instance segmentation models, accurate segmentation of plants and curds was achieved, and plant canopy width, leaf area, and curd traits were automatically extracted to track their dynamic changes. Evaluation results showed that YOLO12s-seg was the optimal model. It can achieved a segmentation mask mAP 50 of 99.4% for plants and curds in sparsely planted images and showed an advantage in identifying partially occluded early curds beneath inner leaves. Traits such as plant canopy width and curd diameter automatically extracted from segmentation results were highly consistent with manual measurements (R 2 > 0.90). Furthermore, the Richards model and Sine model were used to accurately fit the growth dynamics of leaf area and curd area, respectively. Based on growth kinetics, curds were classified into three types: mature and compact type, peak-burst type, and steady-increase type. Cluster analysis of 47 germplasms based on high-throughput phenotyping data revealed four groups and their growth characteristics: comprehensively coordinated type, mid-maturity compact type, large high-yield type, and curd-dominant type. Integrating the above functions, a platform for cauliflower growth monitoring and phenotypic analysis was developed. It provided full-process support from automatic image processing to growth dynamic analysis. This work provides an effective automated solution for high-throughput phenotyping analysis and growth dynamic monitoring of cauliflower, and offers a referable analytical framework for crop growth pattern research and intelligent breeding decision-making.
Why it matches plant phenotyping methods植物のインスタンスセグメンテーションから葉面積・草冠幅・花蕾形質を自動抽出し、手動測定との技術検証と成長動態解析、統合プラットフォーム開発を行っており、フェノタイピング手法が研究の中心である。
abstractThrough systematic screening and evaluation of instance segmentation models, accurate segmentation of plants and curds was achieved, and plant canopy width, leaf area, and curd traits were automatically extracted to track their dynamic changes.
Rice diseases such as leaf blight, blast, and brown spot remain major constraints on food security and rural livelihoods across Southeast Asia, causing significant yield losses each year. In Indonesia, particularly in Lamongan, East Java, these pathogens threaten smallholder productivity and disrupt national rice supply chains. This study aims to enhance automated rice disease detection under real agricultural conditions by integrating image preprocessing techniques with a deep learning-based detection framework. The main contribution lies in developing a hybrid pipeline that combines RGB-to-grayscale conversion and contrast stretching prior to model training, effectively mitigating low-contrast conditions and noise commonly found in field-acquired image datasets. The enhanced images are subsequently processed using the Faster Region-Based Convolutional Neural Network (Faster R-CNN) with a ResNet-50 backbone to localize and classify disease symptoms. Experiments conducted on a dataset of 1,500 annotated rice leaf images achieved high detection performance, with accuracies of 97.37% for leaf blight, 94.12% for blast, and 95.24% for brown spot. Compared with the baseline Faster R-CNN model, the proposed approach improved classification accuracy from 0.8906 to 0.9297, reduced false negatives from 0.439 to 0.1998, increased foreground classification accuracy from 0.55 to 0.78, and descreased total loss from 0.839 to 0.6493. These results demonstrate that integrating RGB-to-grayscale conversion and contrast stretching significantly enhances feature representation, leading to improved detection accuracy, reduced error rates, and more stable training behavior. Overall, the proposed framework provides a robust and reliable approach for rice disease identification and offers strong potential for practical deployment in precision agriculture systems.
Why it matches plant phenotyping methodsイネ葉の病徴を画像から検出・分類する画像前処理とFaster R-CNNの統合手法を開発し、性能比較・検証しているため、植物表現型取得が中心である。
abstractThe main contribution lies in developing a hybrid pipeline that combines RGB-to-grayscale conversion and contrast stretching prior to model training
Crop attribute detection, as a key component of intelligent agricultural harvesting machinery, plays a crucial role in harvesting efficiency, loss reduction, and autonomous operation control. Compared with existing reviews on artificial intelligence and sensing technologies in agriculture, this review focuses on crop attribute detection scenarios oriented toward the intelligent decision-making and control requirements of agricultural harvesting machinery. It mainly analyzes crop attributes that affect harvesting operations, as well as the sensors and algorithms involved in detecting these attributes, and further clarifies the relationship between detection methods and control decisions in agricultural harvesting machinery. For grain crops, the key attributes relevant to harvesting operations include plant height, plant density, spike number, crop lodging, canopy structure, and crop position. For fruit and vegetable crops, the key attributes relevant to harvesting operations include maturity, position, and quality. From the perspectives of multi-source data acquisition, data analysis, and attribute detection algorithms, the key technologies in the field of crop attribute detection are systematically summarized and analyzed, including sensors used in crop attribute detection, such as RGB, spectral, near-infrared, and LiDAR sensors, as well as data analysis and recognition approaches, such as image classification, object detection, and point cloud analysis. The complexity of field environments and the dynamics of machine operation are analyzed, highlighting the technical bottlenecks of current detection systems in environmental adaptability, real-time responsiveness, and resistance to interference. To address these challenges, feasible optimization directions were proposed, including multi-sensor fusion, weakly supervised learning, and few-shot learning. This review aims to provide systematic references and theoretical support for the coordinated development of crop detection and control decision-making in intelligent agricultural harvesting systems.
Why it matches plant phenotyping methods収穫機械向けではあるが、草丈・密度・穂数・倒伏・群落構造・成熟度など植物の形態・状態を検出するセンサーと解析手法を体系的にレビューしており、表現型取得法が中心である。
titleA Review of Crop Attribute Detection for Agricultural Harvesting Machinery
Abstract Citrus canker and melanose substantially reduce the visual quality and commercial value of orange fruit, yet routine diagnosis remains largely dependent on subjective visual inspection. This study presents an interpretable color–texture–shape feature fusion framework for RGB-based classification of healthy, canker-infected, and melanose-affected orange fruit. Each image was represented by a compact descriptor integrating HSV color histograms, Local Binary Pattern micro-texture features, and Histogram of Oriented Gradients edge–shape information, followed by normalized feature concatenation and one-vs-rest Logistic Regression classification. On a balanced held-out test set, the proposed pipeline achieved 93.93% overall accuracy and a macro-F1 score of approximately 0.94, with class-wise F1-scores of 0.927 for citrus canker, 0.933 for healthy fruit, and 0.958 for melanose. Error analysis showed that residual misclassifications were concentrated mainly along the canker–healthy boundary. These findings demonstrate that well-designed handcrafted descriptors can provide accurate, transparent, and diagnostically meaningful citrus fruit disease recognition.
Why it matches plant phenotyping methodsRGB画像から果実の色・テクスチャ・形状特徴を抽出し、病徴状態を分類する方法が研究の中心であるため、植物病害表現型の画像ベース手法として含める。
abstractThis study presents an interpretable color–texture–shape feature fusion framework for RGB-based classification of healthy, canker-infected, and melanose-affected orange fruit.
Early and reliable detection of rose diseases is important for automating plant monitoring and timely intervention throughout the crop lifecycle. In this context, leaf-image analysis combined with machine learning offers a practical approach for disease detection in roses. This study tests a binary classification framework that distinguishes diseased leaves using convolutional neural networks (CNNs). Three architectures were evaluated: a lightweight CNN trained from scratch as a baseline model, and two residual network models fine-tuned through transfer learning from weights pretrained on a large-scale visual recognition dataset. To assess robustness, two preprocessing strategies were also compared: a lightweight hue-based leaf isolation method that preserves full color information, and a grayscale conversion approach without masking. Experimental results obtained on a small held-out test set show strong classification performance across all evaluated models. At the same time, the findings indicate that additional validation is needed on more diverse datasets to confirm generalization under varying lighting conditions, background complexity, and plant growth stages. The results support the feasibility of CNN-based disease detection for roses and highlight its potential for integration into automated monitoring workflows.
Why it matches plant phenotyping methodsバラ葉画像から病害状態をCNNで推定する分類手法の開発・比較が研究の中心であり、植物の病害表現型を直接評価しているため。
abstractleaf-image analysis combined with machine learning offers a practical approach for disease detection in roses.
Abstract Reliable edge-based citrus disease screening requires more than high benchmark accuracy; it must identify when an image and its prediction are trustworthy under field conditions. This study analyzed a balanced three-class orange image corpus comprising healthy fruit, citrus canker, and melanose using a lightweight HSV–LBP–HOG representation and calibrated confidence profiling. Rather than treating all classifications as equally actionable, the analysis linked prediction confidence with image-quality limitations and residual error patterns. The results showed that difficult cases were concentrated around the early canker–healthy boundary, where weak chromatic changes, subtle rind texture, shadowing, blur, and non-disease surface defects reduced diagnostic reliability. A quality-aware triage scheme was therefore established to separate direct acceptance, image recapture, and expert or local refinement. The findings support a practical field-reliability framework for low-cost orange disease screening on mobile and edge devices.
Why it matches plant phenotyping methods柑橘病害の画像分類について、画像品質、失敗分類、信頼度校正、現場トリアージを中心に評価しており、植物の病徴・病害状態を推定する手法の検証が主題である。
abstractThis study analyzed a balanced three-class orange image corpus comprising healthy fruit, citrus canker, and melanose using a lightweight HSV–LBP–HOG representation and calibrated confidence profiling.
Abstract Edge-based citrus disease screening requires not only accurate recognition but also low latency, modest memory use, robustness to imperfect image acquisition, and flexible decision thresholds for practical field operation. This study evaluated a lightweight classical vision pipeline as an operational screening engine for orange fruit disease detection under CPU-only deployment constraints. RGB images were represented using HSV color histograms, Local Binary Pattern texture descriptors, and Histogram of Oriented Gradients shape features, followed by one-vs-rest Logistic Regression classification. Beyond classification accuracy, the system was assessed through computational profiling, perturbation robustness, descriptor-level accuracy–latency trade-offs, and threshold-tunable canker detection. The pipeline achieved macro-F1 ≈ 0.94 while requiring only ~ 0.482 ms/image for feature extraction, negligible classification latency, ~ 106 MB RAM, and ~ 6% CPU utilization. Robustness analysis showed stable performance under brightness, contrast, crop, rotation, and moderate noise perturbations. These findings support classical feature-based vision as a practical, transparent, and resource-efficient edge-screening strategy for sustainable citrus disease monitoring.
Why it matches plant phenotyping methods柑橘果実の病徴・かんきつかんきつ類かいよう病を画像から検出する古典的コンピュータビジョン手法が中心で、精度、頑健性、計算性能、閾値調整を評価しているため。
abstractThis study evaluated a lightweight classical vision pipeline as an operational screening engine for orange fruit disease detection under CPU-only deployment constraints.
Artificial intelligence (AI)-enabled camera sensor systems are increasingly transforming precision agriculture by providing non-destructive, rapid, and scalable methods for monitoring crop health. Two of the most critical applications are the detection of crop water stress and the assessment of pesticide requirement through pest, disease, and symptom recognition. This literature review synthesizes published work on RGB, thermal, multispectral, and hyperspectral imaging integrated with machine learning and deep learning methods for agricultural decision support. The reviewed studies show that thermal and hyperspectral imaging are particularly effective for water stress detection, whereas RGB and multispectral systems are highly practical for identifying disease symptoms, pest infestation, and spray targets. The literature further indicates a shift from simple classification toward real-time decision support, multimodal fusion, explainable AI, and precision input application. This review discusses core sensing technologies, major algorithmic approaches, research findings from key studies, present limitations, and future research directions. Overall, AI camera sensor systems offer substantial potential for reducing water wastage, minimizing excessive pesticide use, and improving sustainable agricultural productivity.
Why it matches plant phenotyping methods作物の水ストレスや病害症状を画像・センサーから推定する手法を中心に整理したレビューであり、植物状態の取得・推定方法が中核です。
abstractThis literature review synthesizes published work on RGB, thermal, multispectral, and hyperspectral imaging integrated with machine learning and deep learning methods for agricultural decision support.
Understanding plant growth dynamics requires imaging across day-and-night cycles to quantify growth, movement and development in the aerial plant body and to capture the rhythmic nature of these processes. This requires imaging in light during the day and in darkness at night without perturbing plant physiology. Nighttime imaging has typically depended on infrared (IR) illumination, producing monochrome datasets that require specialised hardware and separate analysis pipelines when combined with daytime RGB imaging. Here, we evaluated very low-intensity green (dimG) illumination from standard LEDs as a practical alternative for colour-consistent nighttime imaging and assessed its physiological impact in Arabidopsis thaliana and Lactuca sativa (lettuce). We show that high resolution colour images can be obtained under dimG using low- cost cameras, with sufficient consistency between full-spectrum and dimG images to allow direct comparison and unified image analysis. We show that very low-fluence green light (<0.5 μmol m -2 s -1 ) does not sustain circadian oscillations of gene activity under continuous exposure and does not perturb rhythms when applied during the dark phase of diel cycles. DimG imaging enabled accurate detection of diel leaf movement profiles in Arabidopsis circadian mutants, revealing genotype-specific phase differences under varying photoperiods. In lettuce, dimG pulses and continuous dimG enabled accurate quantification of diel leaf movement without affecting growth, stomatal opening, electron transport rate or chlorophyll content. Motion profiles under continuous dimG mirrored those under darkness. Our findings establish dim green illumination as a cost-effective solution for night-time imaging, simplifying phenotyping workflows with minimal impact on physiology.
Why it matches plant phenotyping methods植物の夜間画像取得用の低強度緑色照明を開発・生理影響評価し、葉運動の定量と統合的な画像解析ワークフローを実証しており、フェノタイピング手法が中心です。
abstractHere, we evaluated very low-intensity green (dimG) illumination from standard LEDs as a practical alternative for colour-consistent nighttime imaging and assessed its physiological impact in Arabidopsis thaliana and Lactuca sativa (lettuce).
Latest imaging technologies play a vital role in the extraction of plant phenotypic traits in high ranges. Most existing analytical methods treat these traits as independent features, overlooking the complex interaction patterns that focus on plant responses to environmental stress. Proposed Perceptual Graph Kernel (PGK) framework model address the limitation in terms of image-derived phenotypic traits plat information graph structured interaction networks leverages perceptual similarity learning to capture higher-order phenotypic patterns. In the PGK framework, traits extracted from RGB (Red, Green, Blue) and multispectral imagery are encoded as nodes, and biologically meaningful relationships amongst trait pairs are represented as weighted edges. Extracted trait values are continuously transformed into perceptual states to enhance biological interpretability, and a graph kernel is employed to measure similarity between trait graphs. Experiments performed in an agricultural field with a precision agriculture dataset for plant stress phenotyping demonstrated that the proposed PGK achieved 93.8% classification accuracy, improving performance by 5.3 percentage points over the CNN baseline. The outcome results clearly highlight the effectiveness of the perceptual graph model for plant phenotyping and provide a robust, interpretable computational framework for sustainable crop monitoring and decision-support in precision agriculture.
Why it matches plant phenotyping methods画像由来の植物形質を抽出・関係グラフ化し、ストレス表現型分類を行う計算手法が研究の中心であるため。
abstractProposed Perceptual Graph Kernel (PGK) framework model address the limitation in terms of image-derived phenotypic traits
Reproduction assets foundThe paper's Data Availability Statement points to a public GitHub repository (marathonengineer/Agriproject) containing the datasets used in this plant stress phenotyping study. The other allowed URL (PlantCV) is a generic phenotyping library, not a paper-specific asset.Dataset · publicThe datasets used in this study are available in publicly accessible online repositories. The repository can be accessed at: https://github.com/marathonengineer/Agriproject.Open asset ↗marathonengineer/Agriprojecthtml-lines:589-657Code / dataset availability confirmedEurope PMC · checked 5 Sept 2026
Climate change-induced drought increasingly constrains water management in mixed-species urban gardens, requiring scalable and non-destructive approaches. This study proposes an integrated framework combining chlorophyll fluorescence, RGB image indices, and machine learning to classify plant drought response patterns. Ten garden plant species were evaluated under varying soil moisture conditions. Hierarchical cluster analysis integrating fluorescence parameters and RGB indices identified three physiologically defined response clusters, and their reproducibility using RGB indices alone was assessed. A total of 1,629 samples were augmented to 1,881 using the synthetic minority over-sampling technique (SMOTE) to address class imbalance. A support vector machine (SVM) model with a radial basis function kernel, using green leaf index (GLI), normalized green-red difference index (NGRDI), blue-green pigment index (BGI), and soil moisture (%) as predictors, achieved an accuracy of 0.91 and a Kappa coefficient of 0.84. In contrast, PLS-DA showed lower performance (accuracy 0.79, Kappa 0.65), indicating limited separability under linear assumptions. These results demonstrate that RGB indices combined with nonlinear models were able to reproduce physiologically defined drought response patterns under the given conditions. As a proof of concept, this study demonstrates the potential of the proposed framework; however, its generalizability is limited by the controlled greenhouse setting, the relatively small number of species, and the lack of external validation in heterogeneous field environments. The framework may provide a cost-effective approach for classifying plant drought responses and has the potential to support the grouping of plants with similar water requirements, which could contribute to improved irrigation management in mixed-species gardens under further validation.
Why it matches plant phenotyping methodsRGB画像指標と機械学習により、植物の干ばつ応答パターンという生理状態を分類し、蛍光測定との再現性を評価しているため、表現型取得・抽出手法が中心です。
abstractThis study proposes an integrated framework combining chlorophyll fluorescence, RGB image indices, and machine learning to classify plant drought response patterns.
Reproduction assets foundThe paper explicitly states that the authors' analysis code (data processing, feature extraction, SVM/PLS-DA modeling) is publicly deposited on Zenodo with a DOI matching an allowed URL. The phenotype datasets are only said to be in the manuscript/supplementary files, so the code deposit is the qualifying paperSpecificCode · publicThe code supporting the findings of this study, including data processing, feature extraction, and machine learning modeling is available at Zenodo: https://doi.org/10.5281/zenodo.19127295 .Open asset ↗Zenodo · 10.5281/zenodo.19127295lines:98-116Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
Introduction Addressing the core bottleneck in traditional crop models-the disconnect between morphology and physiological function at the organ scale and their limited dynamic response to environmental changes-this study aimed to construct a multi-source data fusion maize growth model for simultaneous organ-scale simulation. Methods We developed a closed-loop Environment-Driven-Functional Response-Morphological Feedback (EDFM) architecture. By integrating environmental time-series data, RGB images, and 3D point clouds, we created a multimodal fusion model based on a gated attention network. This approach adaptively weights multi-source features and pioneers a bidirectional morphology-physiology feedback loop based on physiological development time (PDT) and NURBS surfaces. The WOFOST moisture response function was also improved. Results The model significantly enhanced the simulation accuracy of organ-scale growth, reducing the root mean square error (RMSE) for plant height by 74.6% through a morphology-physiology dynamic weighting mechanism. More fundamentally, it resolved the disconnect between morphological and physiological processes. The improved plant height prediction validates the model's effectiveness at the organ scale. Discussion The pioneering "physiology-morphology" parallel simulation architecture provides an interpretable theoretical model and robust quantitative tools for designing high-photosynthetic-efficiency plant architecture and enabling precision water-fertilizer management.
Why it matches plant phenotyping methodsRGB画像・3D点群・環境データを統合し、器官スケールの形態と生長をシミュレーションする手法を開発しており、植物形質（草丈など）の推定が中心的な技術貢献である。
abstractWe developed a closed-loop Environment-Driven-Functional Response-Morphological Feedback (EDFM) architecture.
The transition of coal regions under the European Green Deal and Just Transition Fund creates a need for quantitative, transparent monitoring of ecological recovery on post-mining land. This study presents an autonomous UAV-based methodology for high-resolution monitoring of vegetation dynamics on a reclaimed coal waste heap in Upper Silesia, Poland. A DJI Mavic 3 Multispectral platform with RTK positioning conducted approximately biweekly flights from August 2024 to October 2025 over three study plots acquiring RGB and multispectral imagery at approximately 4 cm/pixel. Photogrammetric processing in DJI Terra produced radiometrically corrected orthomosaics and NDVI maps, which were analyzed using an automated QGIS workflow for reprojection, clipping, NDVI-based classification, and quantification of vegetation area across three different reclamation variants. The results indicate that intensive soil conditioning through the application of compost derived from bio-waste achieved a maximum vegetation cover of 94.4%. This treatment consistently maintained the highest level of cover during periods of environmental stress and significantly surpassed both seeding-only treatments and those combining seeding with irrigation. Baseline vegetation cover below 6% confirmed the necessity of active reclamation. This workflow provides rapid and reproducible metrics that are suitable for adaptive management and regulatory reporting. It also offers a scalable template for monitoring coal waste heaps across Europe undergoing SDG-aligned reclamation.
Why it matches plant phenotyping methodsUAVマルチスペクトル画像と自動QGISワークフローにより、植生被覆を plot レベルで定量化する再現可能な計測手法が研究の中心であるため、植物状態の画像ベース表現型計測として採用。
abstractThis study presents an autonomous UAV-based methodology for high-resolution monitoring of vegetation dynamics on a reclaimed coal waste heap
The leaf area index (LAI) is a key parameter for characterizing crop growth and water use efficiency. Therefore, efficient and accurate monitoring of LAI is essential for precision rice management. To overcome the limitations of traditional LAI measurement methods, which are time consuming, labor intensive, and difficult to scale, this study proposes an inversion framework that integrates multi-source UAV remote sensing features with machine learning models. The framework incorporates color indices (CIs) derived from RGB imagery, vegetation indices (VIs) derived from multispectral data, texture features (TIs), and texture feature indices (TFIs), and employs six machine learning algorithms to develop optimized LAI estimation models for the rice booting stage. The results indicate that at a flight altitude of 30 m, the CNN model integrating CIs and TIs achieved an accuracy of R 2 = 0.815. At 60 m, the RF model combining VIs and TFIs showed superior performance, with an R 2 of 0.866. Further integration of CIs, VIs, and TFIs at 30 m produced the best results, increasing R 2 to 0.901, reducing RMSE to 0.273, and raising RPD to above 3.0. These findings demonstrate that TFIs significantly enhance the spectral-spatial representation capability of multispectral data, thereby improving model accuracy. The combined use of CIs and VIs across different sensors compensates for the inherent limitations between spectral and spatial information, while the integration of multi-resolution TIs and TFIs effectively overcomes the constraints of single-source data. Overall, the proposed approach provides a robust and efficient solution for high-precision LAI estimation during critical growth stages of rice, offering strong support for precision agricultural management.
Why it matches plant phenotyping methodsUAV画像・マルチスペクトル特徴量と機械学習によるイネLAI推定フレームワークの開発・性能評価が研究の中心であり、植物形質の取得手法に該当する。
abstractthis study proposes an inversion framework that integrates multi-source UAV remote sensing features with machine learning models.
Modern agriculture must balance productivity with sustainability. In this context, unmanned aerial vehicles (UAVs) offer flexible, cost-effective tools for crop and soil monitoring in precision agriculture. This study aimed to evaluate the potential of UAV-borne RGB imagery, combined with vegetation indices and machine learning, to estimate surface soil properties and crop physiological traits in peanut (Arachis hypogaea) cultivation. A factorial field experiment with four varieties, two planting densities, and two tillage systems was monitored using high-resolution RGB orthomosaics acquired at key phenological stages. From these images, 17 RGB-based indices were computed and related to soil variables and crop traits using Spearman correlation and two regression algorithms: Random Forest (RF) and k-Nearest Neighbors (KNN). RF models outperformed KNN, with the Red Chromatic Coordinate (RCC) index achieving an R2 of 0.87 for predicting soil organic matter content. Indices such as visible NDVI and the Green Vegetation Index also provided robust estimates of canopy condition and leaf chlorophyll. Overall, the results demonstrate that UAV RGB imagery, processed through simple vegetation indices and RF models, constitutes an effective, low-cost approach for monitoring key agronomic parameters in peanut farming.
Why it matches plant phenotyping methodsUAV RGB画像と機械学習による作物生理形質・キャノピー状態・葉緑素の推定手法を中心に評価しており、植物表現型取得・推定が実質的な貢献である。
abstractThis study aimed to evaluate the potential of UAV-borne RGB imagery, combined with vegetation indices and machine learning, to estimate surface soil properties and crop physiological traits in peanut (Arachis hypogaea) cultivation.
Accurate plant segmentation in thermal imagery remains a significant challenge for high throughput field phenotyping, particularly in outdoor environments where low contrast between plants and weeds and frequent occlusions hinder performance. To address this, we present a framework that leverages synthetic RGB imagery, a limited set of real annotations, and GAN-based cross-modality alignment to enhance semantic segmentation in thermal images. We trained models on 1128 synthetic images containing complex mixtures of crop and weed plants in order to generate image segmentation masks for crop and weed plants. We additionally evaluated the benefit of integrating as few as 20 real, manually segmented field images within the training process using various sampling strategies. When combining all the synthetic images with a few labeled real images, we observed a maximum relative improvement of the mean IoU score of 25% compared to the synthetic-only baseline. Cross-modal alignment was enabled by translating RGB to thermal using CycleGAN-Turbo, allowing robust template matching without calibration. Results demonstrated that combining synthetic data with limited manual annotations and cross-domain translation via generative models can significantly boost segmentation performance in complex field environments for multi-model imagery.
Why it matches plant phenotyping methods熱画像における作物・雑草の分割を対象とし、合成データ、少数の実画像、GANによるモダリティ間整合を用いた高スループット圃場フェノタイピング手法を開発・評価しているため。
abstractAccurate plant segmentation in thermal imagery remains a significant challenge for high throughput field phenotyping
Reproduction assets foundThe paper's real annotated cowpea segmentation images and its synthetic Helios-generated training imagery are both publicly available on Hugging Face per the Data Availability statement. No author analysis code repository with explicit deposit language is provided (Helios and AgML are generic third-party tools, not theDataset · publicendix A
Supplementary data to this article can be found online at https://doi.org/10.1016/j.plaphe.2026.100214 .
Appendix A.
Supplementary data
The following is the Supplementary data to this article:
Multimedia component 1
Data availability
Data can currently be accessed through Huggingface [ 75 ]. The real data is found here: https://huggingface.co/datasets/earlranario/cowpea-segmentation . The synthetic data is found here: https://huggingface.co/datasets/earlranario/cowpea-synthetic-segmentation .Open asset ↗earlranario/cowpea-segmentationlines:341-366Plant phenotyping relevance match · UnverifiedOpenAlex · checked 15 Sept 2026
Published1 May 2026Institutional Repositories DataBase (IRDB)
Nitrogen (N) management is critical for optimizing growth and fruit quality in open-field strawberry cultivation, demanding advanced technological solutions for reliable nutrient assessment. However, visual symptom diagnosis, though widely utilized for nutrient monitoring, is inherently subjective and prone to observer bias, resulting in inconsistent and often unreliable assessments. While available accurate tissue analysis is destructive and costly. Nondestructive, in-field imaging techniques such as the normalized difference vegetation index (NDVI) exist but require expensive multispectral imaging systems. To address these limitations, this study developed a streamlined methodology for in-field N status detection using deep learning on standard RGB images. The experiment utilized 'Yotsuboshi' strawberries in a randomized complete block design with sufficient nitrogen (T1) and deficient nitrogen (T2) treatments. To mitigate ambient light variability, a key challenge in open-field phenotyping, a low-cost phenotyping cylinder was developed for standardized smartphone image acquisition. Rigorous four-stage annotation criteria were also introduced to classify the nitrogen status in strawberry leaves as NormalN, LowN, or AdvancedLowN, ensuring a high-quality novel dataset. A YOLO11 model trained on this dataset achieved precision, recall, and mAP50 values exceeding 99%. Subsequent testing using the phenotyping cylinder yielded a mAP50 of 87%. In-field validation without a phenotyping cylinder also demonstrated robust performance under diffuse cloudy conditions (82.7% mAP50), outperforming direct sunlight (79% mAP50). Moreover, the model's classifications of 'NormalN' and 'LowN' statuses strongly corresponded with NDVI measurements, validating the accuracy of the RGB-based approach. This research demonstrates the significant potential of combining deep learning and phenotyping cylinder to create a rapid, low-cost, nondestructive and reliable tool for in-field nitrogen detection, with possible application across different crops and environmental conditions.
Why it matches plant phenotyping methodsイチゴ葉の窒素状態という植物状態を、RGB画像・深層学習・低コスト撮影筒で非破壊推定する手法を開発し、データセット作成と複数条件で検証しており、表現型取得・抽出法が研究の中心である。
abstractthis study developed a streamlined methodology for in-field N status detection using deep learning on standard RGB images.
Abstract This study assesses high‐throughput red, green, and blue (RGB) imaging as an approach for detecting subtle phenotypic differences under well‐watered and reduced watering conditions in genetically uniform wheat ( Triticum aestivum L.) populations. It aims to support the design of breeding populations by identifying parents with complementary coping mechanisms that can be combined in crosses to produce superior progeny. We used RGB imaging to monitor side‐projected area (SPA) in BC 2 F 6 wheat progenies under well‐watered, pre‐anthesis, and post‐anthesis reduced watering conditions. SPA was modeled with logistic growth curves per genotype to extract dynamic canopy traits, which, together with the area under the SPA‐based growth curve, were then correlated with yield, straw biomass, harvest index, and spike traits measured at maturity. Despite genetic similarity, RGB‐based imaging revealed distinct phenotypes under normal conditions and stress response strategies among wheat lines, highlighting the value of dynamic, non‐destructive phenotyping for identifying complementary response patterns. Under well‐watered conditions ( n = 36), area under the curve was strongly associated with grain weight ( R 2 = 0.76, 95% confidence interval [CI]: 0.59–0.87), but relationships weakened under reduced watering, especially post‐anthesis, indicating a reduced association of canopy size with reproductive output. The data revealed contrasting response patterns among breeding lines based on characteristics of the logistic growth curve under normal conditions, their recovery slope after pre‐flowering reduced watering, or conversion of their straw biomass into harvestable grains. RGB imaging enables real‐time, non‐destructive detection of reduced watering responses in genetically similar wheat lines and provides complementary in‐season data to design next‐generation breeding populations for climate‐resilient cultivars.
Why it matches plant phenotyping methodsRGB画像で動的なキャノピー形質を抽出し、育種利用に向けた非破壊・リアルタイム表現型解析を実質的に評価しているため。
abstractThis study assesses high‐throughput red, green, and blue (RGB) imaging as an approach for detecting subtle phenotypic differences
Abstract Fusarium head blight (FHB) of wheat ( Triticum aestivum L.) is primarily caused by the fungal pathogen Fusarium graminearum . This disease can cause significant economic loss due to decreasing yield, reducing seed quality, and the production of deoxynivalenol (DON); therefore, resistance to the disease is a primary concern for breeders. Phenotyping methods largely depend on the resistance mechanism being evaluated, but traditional approaches are often time‐consuming, subjective, and largely inaccurate. This review explores and contrasts traditional and modern phenotypic methods for various FHB resistance components. Digital image‐based phenotyping spans low‐cost RGB (red, green, blue) (i.e., Bayer array) sensors, multispectral sensors, and hyperspectral sensors. Across these sensors, approaches using spectral indices or deep learning have shown strong promise for detecting and classifying infection in both wheat spikes and kernels. Hyperspectral imaging has been largely explored and can be used to accurately estimate infection in spikes and kernels, as well as estimate DON content in the grain, using spectral indices or models input with specific wavebands. However, waveband‐specific approaches do not generalize well to new data, and hyperspectral imaging is significantly more resource‐intensive than RGB or multispectral cameras, limiting its practicality for most breeding programs. Phenotypic approaches using spectral indices and/or deep learning on digital images show the most potential for use in wheat breeding, due to their scalability and low cost. However, the widespread adoption of these techniques will depend on standardized imaging protocols, robust generalization across diverse genotypes, and effective integration into breeding pipelines.
Why it matches plant phenotyping methodsコムギ赤かび病抵抗性の表現型取得手法を、従来法からRGB・マルチスペクトル・ハイパースペクトル画像解析まで比較・レビューしており、フェノタイピング手法が中心である。
abstractThis review explores and contrasts traditional and modern phenotypic methods for various FHB resistance components.
Tea diseases, including brown and gray blight, result in significant yield and quality losses, especially in Longjing tea production. Traditional detection methods are prone to errors, while existing deep learning models often struggle to be robust under natural field conditions. To address these challenges, an improved lightweight detection model, asymmetric multi-level (AML) mechanism, dynamic snake convolution (DSC), and scalable intersection over union (SIoU) loss function-You Only Look Once (YOLO) (ADS-YOLO), was developed and validated. In the method, a dataset comprising 5694 smartphone-captured images of tea leaves was established under natural lighting. Enhancements were implemented in the YOLO11n baseline algorithm through incorporation of the SIoU loss function for better bounding box regression, DSC, which realizes adaptive feature extraction based on the dynamic spatial context, and an AML mechanism, which achieves lightweight feature fusion via adaptive multi-scale design. The results showed that ADS-YOLO achieved a precision of 0.935 and a recall of 0.870, compared to 0.894 and 0.818, respectively, when the baseline YOLO11n was used. Importantly, ADS-YOLO demonstrated a real-time performance of 137.1 frames per second (FPS), coupled with reduced computational costs. ADS-YOLO improved the mean average precision (mAP) at intersection over union threshold of 0.5 (mAP@0.5) by 6.4% compared with YOLOv5n and achieved up to 44.6% higher accuracy than YOLOv7t. In conclusion, ADS-YOLO achieved high accuracy, providing a scalable solution for real-time crop health monitoring and sustainable precision agriculture for tea production.
Why it matches plant phenotyping methods茶葉の病害症状を画像から検出する深層学習手法を開発・検証しており、植物の病害状態の取得が研究の中心である。
abstractan improved lightweight detection model, asymmetric multi-level (AML) mechanism, dynamic snake convolution (DSC), and scalable intersection over union (SIoU) loss function-You Only Look Once (YOLO) (ADS-YOLO), was developed and validated.
ABSTRACT Lodging is a major contributor to decreased yield in tef, a staple cereal crop in Ethiopia. Semidwarf varieties have been developed with a goal to increase yield through reduced lodging, but studying lodging susceptibility currently requires a labor‐intensive, imprecise, manual scoring method. Here we present workflows for analyzing tef stand height from UAS sensors across time to both predict lodging later in the season with early height and to measure the severity of lodging after a storm event. We compare 3D point clouds generated by photogrammetry from RGB images with those generated from LiDAR to estimate height, demonstrating that they produce similar results, despite differences in cost. Stand height and lodging can both be accurately measured with low‐cost UAS, reducing the need for manual measurements and increasing precision and temporal resolution in plant breeding programs.
Why it matches plant phenotyping methodsUAS画像・LiDARによるテフの草高と倒伏程度の推定ワークフローを開発・比較し、育種での測定精度向上を示す中心的な表現型計測研究。
abstractHere we present workflows for analyzing tef stand height from UAS sensors across time to both predict lodging later in the season with early height and to measure the severity of lodging after a storm event.
Reproduction assets foundThe paper's Data Availability Statement and Methods sections point to a public GitHub repository containing the authors' analysis code and associated data (including PheNode sensor data), plus the PlantCV-Geospatial package used for the RGB/LiDAR height and lodging analysis.Code · publicthe USDA NIFA AFRI (Grant Number
2022-
67021-
36467 to N.F.), and by the Bellwether Foundation.
Conflicts of Interest
Getu Beyene has patent “Lodging resistance in Eragrostis tef” pending
to Donald Danforth Plant Science Center.
Data Availability Statement
Code and data associated with this manuscript are available on GitHub
(https://github.com/danforthcenter/teff-manuscript).References
Abebe, Y., A. Bogale, K. Michael Hambidge, B. J. Stoecker, and R. S.
Gibson. 2007. “Phytate, Zinc, Iron and Calcium Content of Selected Raw
and Prepared Foods Consumed in Rural Sidama, Southern Ethiopia,
and Implications for Bioavailability.” Journal of Food Composition and
Analysis 20, no. 3: 161–168.
AssOpen asset ↗danforthcenter/teff-manuscriptpdf-raw-page:8 lines:1-98Code · publicyzing images of plants (Gehan
et al. 2017; Schuhl et al. 2026) that provides a framework for
measuring and storing observations extracted per object within
each image. All code associated with these analyses is available
on GitHub (https://github.com/danforthcenter/teff-manuscript),
as well as the PlantCV-
Geospatial package (https://github.com/danforthcenter/plantcv-geospatial). As observed in the ortho-
mosaic (Figure 1A), tef plots were planted under power lines in
the field, which could not be flown under due to UAS safety re-
strictions. Pixels belonging to powerlines needed to be removed
to measure plot heights. During import, PlantCV-
Geospatial
was used with a height percentile tOpen asset ↗danforthcenter/plantcv-geospatialpdf-raw-page:4 lines:1-107Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
Accurate rapeseed yield and biomass estimation at the meter scale prior to harvest is crucial for precision harvesting. However, there is a scarcity of structured research on the estimation of rapeseed biomass yield. This study aims to address this gap by focusing on rapeseed in Jiangsu Province. Multispectral and RGB images captured by unmanned aerial vehicles (UAVs) were taken during key growth stages (budding, flowering, and podding stages). Using the extracted multidimensional features, we developed biomass-yield estimation models using four machine learning techniques. Subsequently, we employed ensemble learning with multidimensional, multi-stage data and used Shapley additive explanation (SHAP) for feature contribution analysis, thereby constructing a framework for predicting rapeseed harvest characteristics with high estimation accuracy and interpretability. Our analysis indicates that spectral‒texture is the most effective feature combination for biomass estimation, whereas the optimal combination for yield estimation includes three-dimensional (3D) spectral‒textural‒structural features. The synergy of these features, coupled with an ensemble learning model, significantly enhanced the accuracy of rapeseed biomass-yield estimation (biomass: coefficient of determination ( R 2 )=0.72, relative root mean square error (rRMSE)=14.35%; yield: R 2 =0.68, rRMSE=13.67%). The proposed model also achieved stable prediction results across the variety‒density interaction. Overall, this study presents an accurate and generalizable approach for estimating rapeseed biomass yield across various planting patterns, offering new insights for precision harvesting.
Why it matches plant phenotyping methodsUAV画像から抽出した多次元特徴とアンサンブル学習により、ナタネのバイオマス・収量を推定する方法が研究の中心であり、精度評価も実施しているため、植物フェノタイピング手法として適格です。
abstractUsing the extracted multidimensional features, we developed biomass-yield estimation models using four machine learning techniques.
Water stress is a global challenge that severely impacts crop production by hindering essential physiological processes. To address this issue, proximal sensing has emerged as a promising technique for the early identification of stress in vegetables, enabling timely management interventions and optimizing yield. This study aimed to use RGB image indices and an artificial neural network (ANN) model to quantify the responses of various plant traits, such as fresh biomass (FB) weight, dry biomass (DB) weight, canopy water content (CWC), relative chlorophyll content (SPAD), soil moisture content (SMC), and tomato yield across different irrigation levels. Field experiments were conducted during the 2022 and 2023 growing seasons, capturing digital RGB images and measuring plant traits at the flowering and fruit-ripening stages. The results revealed that a reduced irrigation level led to a decrease in various plant traits. The study also revealed significant differences in RGB image indices between different irrigation levels, with strong positive relationships identified for the majority of RGB image indices incorporating green components (G) and R2 reaching 0.99 for various plant traits. However, the red-blue simple ratio (RB) index, which does not consider the G, did not significantly correlate with any of the plant traits. The ANN models achieved high prediction accuracy, with high R2 values reaching 0.99 for various plant traits and yields. These findings underscore the practicality and reliability of employing RGB imaging indices in conjunction with ANN models for effectively managing tomato crop growth and production, particularly under limited water conditions.
Why it matches plant phenotyping methodsRGB画像指標とANNによる植物形質・収量の定量推定が研究の中心であり、予測精度も評価しているため、画像ベース形質推定の方法適用・検証に該当する。
abstractThis study aimed to use RGB image indices and an artificial neural network (ANN) model to quantify the responses of various plant traits
This paper addresses the problem of automated segmentation of plant green biomass in field crop images aimed at improving the accuracy of crop and weed identification. To construct a training dataset for neural network models, an automatic annotation algorithm is proposed, enabling the generation of polygonal object masks without human intervention. The method is based on adaptive analysis of color characteristics of plant fragments with iterative narrowing of the hue range in the HSV color space, combined with an integral quality metric that accounts for the dynamics of contour area and shape. The proposed method achieved an IoU of 93.22% and a DSC of 96.30%, demonstrating a high level of agreement between automatic and manual annotations. The generated masks are used to train segmentation models of the YOLO11-seg family. Models of different scales (n, s, m, l, x) were trained and evaluated using standard metrics, including Intersection over Union (IoU), mAP@0.5, mAP@0.5–0.95, F1-score, and Precision–Recall (PR) curves. Experimental results demonstrate that models trained on automatically generated annotations achieve stable segmentation performance of plant green biomass. The best results were obtained with the YOLO11m-seg model, achieving an F1-score of 0. 772. The results confirm the effectiveness of the proposed approach and demonstrate acceptable segmentation quality, supported by both quantitative metrics and visual analysis. The developed automatic annotation algorithm can be used to expand training datasets in computer vision tasks for agricultural applications.
Why it matches plant phenotyping methods植物の緑色バイオマスを画像から自動抽出するポリゴン注釈法を開発し、手動注釈との一致度で検証しているため、植物表現型取得・抽出法が中心である。
abstractan automatic annotation algorithm is proposed, enabling the generation of polygonal object masks without human intervention
Reproduction assets foundThe authors publicly released the paper-specific generated dataset of polygonal segmentation annotations (masks and supporting materials) on Hugging Face. CVAT is only a generic annotation tool, and no author analysis code or trained model checkpoints are explicitly deposited.Dataset · publicThe generated dataset with polygonal segmentation annotations of crop and weed plants, produced using the proposed algorithm and based on the LincolnBeet Dataset, is publicly available on Hugging Face at: https://huggingface.co/datasets/ivliev123/polygonal_marking_plant_objectsOpen asset ↗Hugging Face · ivliev123/polygonal_marking_plant_objectshtml-lines:438-462Plant phenotyping relevance match · UnverifiedEurope PMC · OpenAlex · checked 5 Sept 2026
To address the low efficiency of manual inspection for corn sowing quality, which is labor-intensive and time-consuming, this study proposes an automatic seedling-stage plant-spacing measurement method based on three-dimensional machine vision. A high-clearance mobile platform equipped with a ZED 2i stereo camera and an industrial computer was developed to acquire RGB images and depth information of corn seedlings in the field in real time. Using the YOLOv11-Pose model, plant keypoints were detected and localized; combined with camera calibration and 3D reconstruction techniques, inter-plant distances were computed automatically. A sowing-quality evaluation framework was then established to enable automated analysis of the Quality of Feed Index (QFI), Multiple Index (MUL), Miss Index (MI), and coefficient of variation. Experimental results indicate that the system operates effectively under three preset plant spacings (15 cm, 20 cm, and 25 cm), achieving a keypoint-detection mAP@0.5 of 0.990 and an mAP@0.5:0.95 of 0.989. In sowing-quality evaluation, the qualified indices produced by the system were 77.83%, 80.36%, and 82.46%, respectively, and the Quality of Feed Index (QFI), Multiple Index (MUL), Miss Index (MI), and coefficient of variation showed trends consistent with manual measurements. The proposed method enables efficient, nondestructive detection of seedling-stage plant spacing and sowing quality, providing reliable technical support for precision sowing and field management.
Why it matches plant phenotyping methods3Dマシンビジョン、姿勢検出、校正、3D再構成を組み合わせ、トウモロコシ個体間距離を自動推定・検証する手法が研究の中心である。
abstractthis study proposes an automatic seedling-stage plant-spacing measurement method based on three-dimensional machine vision.
The article presents the developed algorithm and software for automated monitoring of strawberry plant growth using neural network technologies. The YOLO11x and YOLOx-seg models, pre-trained by transfer learning, are adapted to recognize and classify plants (plant class), leaves (leaf class), and a reference marker (ref_obj class) of a known size. Segmentation of strawberry leaves using the YOLO11x-seg model makes it possible to analyze the morphometric parameters of individual leaf plates (area, perimeter, roundness, aspect ratio). A set of RGB images (2000 pieces) obtained using a GoPro HERO11 camera under controlled laboratory conditions was formed and annotated, followed by augmentation to increase the model's resistance to variations in shooting conditions. The developed algorithm converts the coordinates of the bounding boxes and segmentation masks of recognized objects into metric units using calibration coefficients calculated from a marker of known size (100×100 mm). The software implemented using PyQt5, TensorFlow, Keras, and OpenCV libraries provides not only visualization of results but also data storage in a local SQLite database with the ability to export to JSON and Excel formats. Validation of the model showed high accuracy in detecting plant bounding boxes (mAP50 = 0.906) and leaf segmentation (mAP50 -mask = 0.625). The average processing speed was 20.3 ms/frame for detection and 34.5 ms/frame for segmentation. The measurement error was less than 3.5 % for the overall parameters of the plant and 5.2 % for the morphometric parameters of the leaves, confirming the effectiveness of the method for assessing the height, width and area of plants, as well as the analysis of the leaf apparatus. The research results show the promise of an approach for automating plant phenotyping in real time.
Why it matches plant phenotyping methods植物の成長・葉形態を画像から自動抽出するアルゴリズム、ソフトウェア、データセットを開発し、精度・処理速度・測定誤差を検証しているため、植物フェノタイピング手法が中心である。
abstractThe article presents the developed algorithm and software for automated monitoring of strawberry plant growth using neural network technologies.
Plant stress monitoring is invaluable in realizing sustainable agriculture because it enables the people practicing it to take early measures to counteract losses in yield caused by environmental stressors like drought and nutrient deficiencies, as well as caused by pathogen infections. The proposed study presents a new Multi-Modal Vision Transformer (MMViT) architecture that is designed to combine both thermal and RGB imagery to take detection to the next level. In order to support this methodology and further studies, we are now publicly releasing a new collection of synchronized thermo-RGB image pairs of stressed and healthy plants, collected both in controlled settings and in the field. The data is labeled to differentiate various stress phenotype and contains over 4286 of images, and hence forms a substantial platform to evaluate multimodal plant phenotyping methods. Empirical evaluations indicate that MMViT model achieves a general classification of 94.3% when using the two modalities, which is better than the single-modality ViT used on the thermal images (85.5%) and the RGB images (93.3%). These experimental results emphasize the performance of multimodal fusion whereby the other spectral cues are used to complement a stress classification. The described framework, together with the useful dataset, will contribute to the advancement of precision agriculture as it is an open and data-driven instrument to monitor plant health automatically.
Why it matches plant phenotyping methods熱画像とRGB画像を統合して植物ストレス表現型を分類するモデルを開発し、公開データセットと性能評価も提示しており、植物フェノタイピング手法が中心である。
abstractThe proposed study presents a new Multi-Modal Vision Transformer (MMViT) architecture that is designed to combine both thermal and RGB imagery to take detection to the next level.
Accurate yield estimation is vital for precision wheat management and breeding. Traditional methods based on single growth stages or single-source data cannot capture cumulative growth effects, limiting prediction accuracy. UAV remote sensing provides high-resolution, multi-source, and multi-temporal data, enabling improved non-destructive yield estimation. In this study, UAV-based multispectral and RGB imagery were collected at six key growth stages, and vegetation indices, texture, and color features were extracted to develop yield prediction models using RF, XGBoost, and KNN under single- and multi-temporal scenarios. The results showed that red-edge-based vegetation indices were highly sensitive to wheat yield and outperformed texture- and color-based features. Multi-feature fusion further improved prediction accuracy at key growth stages, particularly during booting and flowering (R 2 = 0.53-0.67). Compared with single-temporal models, multi-temporal data fusion significantly enhanced yield estimation accuracy, achieving a maximum R 2 of 0.72 by integrating data from the late-jointing, booting and flowering stages. Among the algorithms, XGBoost and KNN exhibited superior accuracy and stability across most growth stages. Overall, these results demonstrate that integrating UAV-based multi-source and multi-temporal remote sensing data effectively improves the accuracy and robustness of wheat yield estimation, providing valuable technical support for precision agriculture and phenotyping-assisted breeding.
Why it matches plant phenotyping methodsUAV画像から特徴量を抽出し、機械学習でコムギ収量を推定する手法が研究の中心であり、マルチソース・マルチテンポラル統合の技術評価も行っている。
The decline in environmental quality caused by industrial pollution and climate change has weakened the natural resistance of rice plants (Oryza sativa), increasing their susceptibility to various diseases. Conventional disease identification methods that rely on manual observation are often limited by subjectivity and human visual constraints. This study proposes a deep learning–based system for automatic rice leaf disease classification using the You Only Look Once version 8 (YOLOv8) architecture. The model was trained using a publicly available rice leaf image dataset consisting of 6,889 images categorized into eight classes: Bacterial Leaf Blight, Brown Spot, Leaf Blast, Leaf Scald, Sheath Blight, Narrow Brown Leaf Spot, Rice Hispa, and Healthy Rice Leaf. The research methodology includes image pre-processing, data augmentation, dataset splitting, and training using the YOLOv8n-cls model for 50 epochs. Experimental results demonstrate high classification performance with an accuracy of 99.5%, precision of 99%, recall of 98%, and an F1-score of 0.99. The trained model was then deployed into a web-based application that allows users to upload rice leaf images and obtain real-time disease classification results. The proposed system provides a practical tool to support early detection of rice plant diseases and assist farmers in improving crop management in modern agriculture.
Why it matches plant phenotyping methodsイネ葉画像から病害状態を推定するYOLOv8画像解析手法の開発と性能評価が中心であり、植物病害フェノタイピングに該当する。
abstractThis study proposes a deep learning–based system for automatic rice leaf disease classification using the You Only Look Once version 8 (YOLOv8) architecture.
Reproduction assets foundThe paper's rice leaf disease image dataset (6,889 images, eight classes) used for YOLOv8n-cls training is a publicly available Kaggle dataset cited by the authors with an explicit URL. No author code, trained model, or other paper-specific assets are reported.Dataset · publicThe primary dataset was obtained from a
publicly available dataset on Kaggle [16], which provides a
comprehensive collection of rice leaf disease images for
machine learning research.Open asset ↗Kagglepdf-raw-page:3 lines:1-102Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · checked 13 Sept 2026
Abstract 3D Gaussian Splatting (3DGS) [KKLD23] has transformed novel‐view synthesis from RGB images, yet remains restricted to the visible spectrum. Many applications, including agricultural monitoring, rely on multi‐spectral imaging, where spectral camera alignment and scalability pose major challenges. We present MS‐Splatting—a multi‐spectral 3DGS framework enabling unified multi‐view consistent reconstruction and rendering across both visible and invisible spectra. Our key component is a neural color representation that encodes per‐primitive features shared across spectral bands, decoded through a shallow multi‐layer perceptron into spectrum‐specific radiance. By leveraging inter‐band correlations, this formulation enhances detail while reducing memory consumption compared to independent band modeling via per‐channel modeling with spherical harmonics. Our method enables accurate parallax‐free novel‐view vegetation index rendering for plant monitoring and enhances RGB novel view synthesis quality by exploiting details revealed through multi‐spectral bands. Our evaluation demonstrates that MS‐Splatting exceeds the current leading methods in both categories. In addition, we introduce a multi‐spectral dataset from aerial captures covering outdoor environments, specifically designed for evaluating these applications. We will release our code and dataset to facilitate further research. The project page is located at: https://meyerls.github.io/ms_splatting
Why it matches plant phenotyping methodsマルチスペクトル3D再構成法を開発し、植物モニタリング用の植生指数レンダリングを実現することが中心で、評価用データセットも提供している。
abstractWe present MS‐Splatting—a multi‐spectral 3DGS framework enabling unified multi‐view consistent reconstruction and rendering across both visible and invisible spectra.
Phenotyping plant traits using UAV-based multispectral imaging offers a robust and unbiased approach to assessing crop status. With approximately 70% of smallholder farmers in East and Southern Africa cultivating common beans as a key source of food and income, there is a critical need for accurate and timely measurements of crop health and yield to support data-driven management decisions and disease mitigation. Traditional phenotyping methods are labor-intensive, and existing remote sensing and machine learning approaches remain limited. This study presents a comprehensive framework for plot-level assessment of common bean health and yield using time-series RGB and multispectral imagery. Data collected over three growing seasons (2022–2024) were used to extract canopy variables and vegetation indices (VIs) across phenological stages. For yield prediction, traditional machine learning models achieved a root mean squared error (RMSE) of 242.33 kg ha⁻¹ and an R² of 0.66 using an Extra Trees Regressor. A novel BY-GRU architecture improved performance, achieving an RMSE of 242.40 kg ha⁻¹ and an R² of 0.79. The analysis also identified 45–60 days after sowing as the optimal window for prediction. To address limitations in conventional plant health assessments, this study introduces a novel Health Index. Comparative analysis demonstrated its robustness across genotypes and stronger correlation with yield. Machine learning and deep learning models, including MaxViT, were applied to estimate the Health Index, achieving improved predictive performance. Overall, this work integrates UAV sensing and modelling to provide scalable tools for phenomics, crop management, and breeding.
Why it matches plant phenotyping methodsUAVのRGB・マルチスペクトル画像から作物の健康状態、収量、キャノピー形質を推定するセンシング・機械学習フレームワークが研究の中心であり、植物フェノタイピング手法として適格です。
abstractThis study presents a comprehensive framework for plot-level assessment of common bean health and yield using time-series RGB and multispectral imagery.
Reproduction assets foundThe preprint's DATA AVAILABILITY section states that all processed data required to reproduce the results are publicly available in a Google Drive repository, which qualifies as a paper-specific public phenotype dataset asset. No author analysis code or trained model checkpoints are explicitly deposited.Dataset · publicCommon Bean Breeding Program for
facilitating field trials. We also thank the Phenomics team for their valuable assistance with UAV-based
data collection.
CONFLICT OF INTEREST
The authors declare no conflict of interest.
DATA AVAILABILITY
The datasets generated and/or analyzed during the current study are publicly available at:
https://drive.google.com/drive/folders/1fN3Q9n3bK_YoXFK8VFKZ3uEb13y9iRWj?usp=sharing. This
repository includes all processed data required to reproduce the results presented in this study.
SUPPLEMENTAL MATERIAL
Supp. Figure 1. Drone-based field view of the bean trial site at CIAT Palmira Research Station: A) RGB
image and B) NDVI image.
Supp. Figure 2. Drone Features Open asset ↗pdf-raw-page:40 lines:1-46Code / dataset availability confirmedOpenAlex · Crossref · Europe PMC · checked 5 Sept 2026
Effective high-throughput phenotyping is crucial for modern plant breeding, yet the optimal image acquisition parameters for UAV-based systems in forage crops remain poorly defined. We optimized UAV-based phenotyping methods for a Megathyrsus maximus biparental population, examining how ground sampling distance (GSD), environment, and harvest date affect the accuracy of RGB-derived digital traits in predicting yield and canopy height. Machine learning algorithms and mixed model analyses were applied to evaluate predictive power and heritability. Pixel count and Haralick's entropy showed strong correlations with conventional yield measurements, particularly in Environment 2, while most vegetative indices were poor predictors. Integrating machine learning substantially enhanced predictive power for green and dry matter yield (r > 0.80). For canopy height, machine learning models achieved correlations of 0.71 with ground truth measurements despite weak pairwise correlations. Mixed model analysis revealed high broad-sense heritability (0.7 < H 2 < 0.87) for yield traits, pixel count, and entropy, while vegetative indices and canopy height showed greater environmental susceptibility. Moderate GSD resolutions (0.5–1.0 cm) consistently outperformed both very high (0.27 cm) and very low (1.5 cm) resolutions. Coincidence index analysis demonstrated 80% correspondence between top genotypes ranked by pixel count and conventionally measured dry matter yield. This study provides an optimized framework for UAV-based phenotyping in M. maximus , demonstrating that combining advanced digital traits with machine learning accurately predicts key agronomic traits and significantly enhances genotype selection efficiency in forage breeding programs.
Why it matches plant phenotyping methodsUAV画像取得条件、RGBデジタル形質、機械学習による収量・草高推定を最適化・検証する研究であり、植物表現型取得法が中心的です。
abstractWe optimized UAV-based phenotyping methods for a Megathyrsus maximus biparental population, examining how ground sampling distance (GSD), environment, and harvest date affect the accuracy of RGB-derived digital traits in predicting yield and canopy height.
Reproduction assets foundThe paper's data availability statement points to a public Mendeley Data repository containing the study's UAV-derived digital phenotyping and conventional trait datasets. No author analysis code repository is explicitly deposited.Dataset · publicThe datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://data.mendeley.com/datasets/jrrb76x82h/1 .Open asset ↗jrrb76x82h/1lines:435-487Plant phenotyping relevance match · UnverifiedCrossref · checked 5 Sept 2026
Plant disease diagnosis in real-world agricultural environments is challenged by data scarcity, domain shift, privacy constraints, and limited edge-device resources. This paper proposes FMEL-FSDA, a Federated Multimodal Edge Learning framework with Few-Shot Domain Adaptation for robust field-based plant disease recognition. The framework integrates attention-based RGB–text feature fusion, privacy-preserving federated learning, rapid few-shot personalization, and uncertainty-aware inference within an edge-efficient architecture. Federated training enables collaborative learning across distributed farms without sharing raw data, while few-shot adaptation allows fast deployment to new regions using only 1–10 labeled samples per class. Experiments on the PlantWild in-the-wild dataset show that FMEL-FSDA outperforms centralized, federated, and few-shot baselines, achieving 93.78% accuracy, 93.33% F1-score, and 0.97 AUC. The model maintains strong performance under privacy mechanisms such as gradient perturbation and secure aggregation, reduces communication overhead by up to 4×, and supports low-latency edge inference. Uncertainty estimation and Grad-CAM-based explainability further enhance reliability by identifying low-confidence cases and highlighting disease-relevant regions. Overall, FMEL-FSDA offers a scalable, privacy-aware, and field-ready solution for intelligent plant disease diagnosis in precision agriculture.
Why it matches plant phenotyping methods植物病害の画像ベース認識手法を開発し、複数ベースラインとの比較評価と実フィールド性能検証を行っており、病害状態の取得・推定が研究の中心である。
abstractThis paper proposes FMEL-FSDA, a Federated Multimodal Edge Learning framework with Few-Shot Domain Adaptation for robust field-based plant disease recognition.
Abstract Current phytopathological diagnostic systems rely on manual inspections or laboratory analyses, which delay early detection and limit in-field responsiveness. Phytopathogenic fungal diseases pose a persistent threat to food security, directly affecting the productivity of essential crops such as potato ( Solanum tuberosum ) and tomato ( Solanum lycopersicum ) [1]–[5]. Among these diseases, Phytophthora infestans , the causal agent of late blight, is characterized by its high virulence and rapid spread, capable of generating significant losses in short periods when detection occurs too late [3], [4]. To address this issue, a computer vision and deep learning–based system for multistage detection of fungal infections in potato and tomato crops is proposed. The system comprises a convolutional neural network optimized for edge processing and a mobile robotic platform equipped with a manipulator arm for localized treatment application. The developed model was deployed on a Raspberry Pi 4 connected to a 12-MP Raspberry Pi Camera Module 3 NoIR, responsible for acquiring RGB images in the field. The proposed network was compared with reference architectures—ResNet-50, VGG16, MobileNetV2, and Inception-v3—within a four-stage detection pipeline: crop identification, health-state classification, infection diagnosis, and foliar severity estimation. A dataset of 18,200 images obtained from publicly accessible online sources, under diverse lighting and background conditions, was used, partitioned into 70% for training, 20% for validation, and 10% for testing. Preliminary results show an average accuracy in the range of 0.90–0.92, with inference latencies below 60 ms per image, ensuring smooth performance on the Raspberry Pi 4 without requiring cloud connectivity. Additionally, the network demonstrated higher sensitivity to visual variations compared to the baseline models.
Why it matches plant phenotyping methods植物病害の健康状態・感染・葉面重症度を画像から推定するコンピュータビジョン手法を開発・比較検証しており、植物表現型取得が中心である。
abstracta computer vision and deep learning–based system for multistage detection of fungal infections in potato and tomato crops is proposed
Reproduction assets foundThe paper's CNN training data are two publicly available third-party plant-disease image datasets (Mendeley Data and Kaggle Plant Village) with explicit URLs in the Data Availability Statement. The authors' field-test images and experimental records are only available on request, and no analysis code or trained model/сDataset · public10
Network, DOI: 10.17632/tywbtsjrjv.1, available at
https://data.mendeley.com/datasets/tywbtsjrjv/1, and the
Kaggle Plant Village dataset, available at
https://www.kaggle.com/datasets/emmarex/plantdisease.The field-test images and experimental records generated during
the current study during the real-world evaluation of the
embedded-vision system are available from the corresponding
author on reasonable request.
IX. REFERENCOpen asset ↗10.17632/tywbtsjrjv.1pdf-raw-page:11 lines:1-98Dataset · public10
Network, DOI: 10.17632/tywbtsjrjv.1, available at
https://data.mendeley.com/datasets/tywbtsjrjv/1, and the
Kaggle Plant Village dataset, available at
https://www.kaggle.com/datasets/emmarex/plantdisease.The field-test images and experimental records generated during
the current study during the real-world evaluation of the
embedded-vision system are available from the corresponding
author on reasonable request.
IX. REFERENCES
[1] P. W. Crous, A. Y. Rossman, M. C. Aime, W. C. Allen, T.
Burgess, J. Z. Groenewald y L. A. CastlebuOpen asset ↗Kagglepdf-raw-page:11 lines:1-98Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · checked 5 Sept 2026
Conventional approaches to measuring stomatal conductance (gs) and transpiration often rely on instruments that interfere with plant physiology. Porometers, for example, restrict natural leaf movement, apply pressure, and introduce dry airflow that can alter stomatal behaviour, thereby reducing the relevance of such measurements. Prior studies report discrepancies among devices attributable to such interferences (Toro et al. 2019). To minimise artefacts, transpiration should be estimated remotely without physical contact, which theoretically can be achieved via a thermal leaf energy-balance approach that infers gs from leaf temperature, radiative load, and boundary-layer terms. In this study, we combine 3D plant models, light interception models, and thermal imaging to solve the energy-balance equation of individual leaves, estimating transpiration entirely remotely. Approaches to estimate stomatal conductance based on the energy-balance equation were developed recently to aid phenotyping of plantss. Most methods either imposed rapid changes in air humidity to perturb transpiration and, consequently, leaf temperature (Driever et al. 2023), or relied on ‘dry’ and ‘wet’ reference surfaces (as in Leinonen et al. 2006) to compute stress indices (Vialet-Chabrand and Lawson 2020). These methods require reference materials to assess surface temperatures under maximum and zero transpiration, showing the effect of longwave radiation. However, reference-material methods were constrained by heterogeneity in light interception caused by variation in leaf angle and orientation, because reference surfaces could not reorient like real leaves (Zhang et al. 2025). In this study, we addressed this challenge by using thermal imaging and 3D photogrammetry to capture leaf temperature and geometry noninvasively, allowing parameter estimation for each leaf individually. Here, ρ is the density of air (kg m−3), cp is the specific heat capacity of air (J kg−1 K−1) and rHR is the parallel resistance to heat and radiative transfer on the leaf surface (s m−1), s is the slope of the curve relating saturating water vapour pressure to temperature (Pa °C−1). TL and TA are leaf and air temperatures (°C), respectively, δe is air vapour pressure deficit (Pa), γ is the psychrometric constant (Pa K−1) and rva is the boundary layer resistance to water vapour (s m−1) (Supporting Information S2: Equation S1). The net radiative energy Rn in the energy-balance term was obtained from the same 3D light interception model, which integrates measured direct and lateral scattered irradiance (W m−2) (Supplement Material and Methods, File S2). Stomatal conductance gs (m s−1) is the inverse of stomatal resistance rs (s m−1). To experimentally obtain a wide range of gs values, we grew eggplant (Solanum melongena L.) plants in hydroponic units in growth chambers under four sets of environmental conditions (Table 1). Thirty-day-old plants (4–5-leaf stage) were placed on balances (Supplementary Materials and Methods, File S2). Units were sealed with plastic film to minimise evaporation. Mass loss attributable to transpiration was logged automatically every 30 s. To induce short-term changes in stomatal conductance, we imposed an acute osmotic stress by delivering a saline NaCl solution with high electrical conductivity (60 mS cm−1) to the root zone, producing a steep drop in root osmotic potential. This created rapid physiological and morphological responses that altered incident irradiance at the leaves, leaf temperature, and consequently energy balance, stomatal conductance and transpiration. We chose this stressor for operational simplicity. Any perturbation that modifies transpiration dynamics and thus gas exchange could have served our purpose. The total transpiration of a leaf, Et (kg s−1), is the product of the total conductance to water vapour from the mesophyll to the atmosphere, gv (m s−1), calculated from the estimated stomatal resistance rs (s m−1) and the boundary layer conductance gva (m s−1), the difference between water vapour concentration inside the leaf Cvs (dimensionless), and in surrounding air Cva (dimensionless), the leaf area A (m2), and the density of water ρw (kg/m3) (Jones 1992). Estimated stomatal conductance was obtained from leaf energy balance calculation (Equation 1). Boundary-layer conductance was computed from measured wind speed and leaf dimensions (leaf area, length, width) extracted from structure-from-motion 3D reconstructions (Supporting Information S1: Equation S6; Grace et al. 1980). Transpiration was then calculated for each leaf at each thermal 3D imaging time point, and whole-plant transpiration for comparison with gravimetric logs was the sum of all per-leaf estimates. As a non-invasive approach, we evaluated the plausibility or our model derived stomatal conductance (Equation 2) indirectly by comparing calculated and measured whole plant transpiration. We emphasise that this is not a direct validation of gs. Rather, the close agreement between modelled and measured transpiration across the wide range of environmental treatments, both stressed and non-stressed, provides confidence that the inferred gs is realistic. RGB and thermal images acquired before, during, and after stress application enabled dynamic tracking of leaf position and temperature (Supplementary Material and Methods, File S2). As expected, osmotic stress application had immediate effects on morphology and physiology. While control leaves maintained an angle of around 110° throughout, osmotic shock induced immediate turgor loss and drooping in all environments except one (Figure 1A,B). Leaf angles recovered to pre-stress positions within 1 h, indicating adaptation to the osmotic shock and restoration of turgor. Only environment 4 (high light, low air temperature and low humidity) maintained turgor during stress. Angle shifts were most pronounced in older leaves, which drooped and reduced light interception; younger leaves better preserved structure and turgor (Supporting Information S1: Figure S2). These angle changes also altered incident irradiance at the leaf surface (Supporting Information S1: Figure S3). These morphological responses coincided with increases in leaf temperature, consistent with altered water fluxes and stomatal regulation after stress. Across environments, plants showed a uniform rise in leaf temperature following osmotic stress, regardless of initial temperature (Supporting Information S1: Figure S4). This response held across leaf ages, encompassing older (Figure 1C) and younger (Figure 1D) leaves. Stomatal conductance estimated with our method followed the same pattern, dropping rapidly after osmotic shock in both older (Figure 1E) and younger (Figure 1F) leaves (Supporting Information S1: Figure S5). We estimated no stomatal conductance recovery to pre-stress conditions over the time course of stress exposure. Model-estimated and gravimetrically measured transpiration showed identical time courses across all four environmental conditions (Figure 1G–J). Transpiration rates did not recover to the same extent as leaf turgor, indicating long-term effects of the osmotic shock. Across environments and time points, correlation between model estimated and measured whole-plant transpiration was high (Figure 1K). In this study, stomatal conductance (gs) is a model-derived quantity inferred from the same physically constrained framework and model (leaf temperature, boundary-layer conductance and vapour pressure deficit). Since we did not measure gs directly, we cannot validate gs directly. Instead, we used a non-invasive check via transpiration. Model predictions closely tracked measured transpiration across the four controlled environments. This agreement increases confidence that the inferred gs is realistic, while we acknowledge that transpiration agreement alone is not a rigorous validation and cannot fully rule out compensating errors. Our study demonstrated the potential of our approach to estimate transpiration accurately by combining 3D imaging and thermography with physiological modelling without the use of reference materials that imitate real leaves. This remote approach enables simultaneous assessment of morphological and physiological responses to stress, yielding a more integrated view on plant transpiration and gas exchange. In contrast to chamber and porometer measurements or IR methods requiring wet and dry references or calibration plates, our workflow is reference-free. Absorbed shortwave radiation is derived from measured irradiance and a 3D reconstruction of leaf geometry, with no external reference materials. Moreover, remote measurements avoid continuous pressure from clamp-on porometers, permitting long-term observation and capture of rapid stress responses without sustained damage or microclimate artifacts. Further, the approach is not limited by any clamp on sensors and as such enables multi-leaf tracking. Applied to crop canopies, this approach could improve understanding of canopy processes that influence productivity and enable remote estimation of canopy transpiration. Future research could further improve by replacing our strong saline solution stress by gradual soil drying to depict a more realistic and natural stress while testing the approach under long-term conditions. Recent studies indicate that, with rising atmospheric CO2 concentrations, breeding for reduced stomatal conductance could increases WUE without affecting photosynthetic capacity (Srivastava et al. 2024). As such, remote systems for high-throughput plant phenotyping (HTP) are required to scan vast quantities of plants. We see a potential use of our system for such purposes to quickly estimated whole plant and individual leaf transpiration, as initial image capturing is very fast. A large bottleneck in our work was 3D model generation speed and manual extraction of leaf parameters from these 3D models. Both could be streamlined with more automated software, possibly including neural network solutions. The authors have nothing to report. The authors declare no conflict of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Why it matches plant phenotyping methods3D画像、熱画像、光遮断モデル、エネルギーバランスモデルを統合し、葉ごとの蒸散と気孔コンダクタンスを非侵襲的に推定する手法を開発・評価しており、植物表現型取得が研究の中心である。
abstractIn this study, we combine 3D plant models, light interception models, and thermal imaging to solve the energy-balance equation of individual leaves, estimating transpiration entirely remotely.
Accurate and non-destructive acquisition of leaf chlorophyll content (LCC) in cotton plant canopies is of significant importance for real-time monitoring of cotton growth and implementing precise water and nitrogen management in cotton fields. This study utilized UAV-based RGB imagery combined with real-time kinematic (RTK) technology to efficiently and accurately retrieve LCC under different nitrogen application levels in a cotton field, employing 6 machine learning algorithms: Least Absolute Shrinkage and Selection Operator regression (LASSO), Multiple Linear Regression (MLR), Partial Least Squares Regression (PLSR), Random Forest Regression (RFR), Ridge Regression (Ridge), and Support Vector Regression (SVR). Among these, the SVR model demonstrated the best overall performance, with coefficient of determination (R²), root mean square error (RMSE), relative root mean square error (rRMSE), and mean absolute percentage error (MAPE) values of 0.82, 0.14 mg/g, 8.91%, and 6.99% for the training set, and 0.75, 0.14 mg/g, 8.90%, and 7.83% for the testing set, respectively. Furthermore, the SVR model was applied to retrieve LCC pixel-by-pixel from UAV imagery, and pseudo-color rendering techniques were used to generate spatial distribution maps of LCC in the cotton canopy, visually presenting the spatial variability characteristics of LCC within the field. The results indicate that the cotton canopy LCC estimation method based on UAV RGB imagery combined with RTK technology achieves comparable accuracy to the more expensive multispectral and hyperspectral techniques, without a significant reduction in precision. This approach provides an efficient, low-cost, and reliable method for detecting canopy LCC in small-scale cotton fields.
Why it matches plant phenotyping methodsUAV RGB画像とRTK、機械学習を組み合わせ、ワタ群落の葉緑素含量を推定・検証する手法が研究の中心であるため、植物フェノタイピング手法として採用する。
abstractThis study utilized UAV-based RGB imagery combined with real-time kinematic (RTK) technology to efficiently and accurately retrieve LCC under different nitrogen application levels in a cotton field, employing 6 machine learning algorithms
The quality of forest reproductive material is crucial for successful reforestation and afforestation. While physical seed properties like mass are known indicators of quality, the potential of non-destructive, rapid color analysis for predicting germination in coniferous species requires further exploration. This study investigates the relationship between the seed coat color of individual Pinus sylvestris seeds, quantified in RGB (Red, Green, Blue) space using a flatbed scanner, and their subsequent germination in container nurseries. The resulting images were processed using ImageJ software to measure the mean pixel intensity (0–255) for the Red (R), Green (G), and Blue (B) channels from the segmented seed area, following the «seed–culture» passport methodology [Forestry Engineering Journal 14 | 55 (2024), 37–60]. From a population of individually tracked seeds, we compared the RGB values of germinated (N = 942) and non-germinated (N = 258) seeds after 30 days. Results from the Kolmogorov-Smirnov test showed that non-germinated seeds had significantly lower individual mass (p = 0.0045) and significantly higher pixel brightness values in the R-, G-, and B-channels (p < 0.0001) compared to germinated seeds. Normalized RGB indices also showed significant differences between groups. Our findings demonstrate that seeds with a lighter, more reflective epidermis – indicative of higher RGB brightness – are statistically associated with a lower probability of successful germination under container nursery conditions. This non-destructive, low-cost method shows significant promise for the rapid pre-sorting of Scots pine seeds. It offers a practical tool to improve the efficiency and predictability of seedling production in forest nurseries by increasing the proportion of viable seeds in sowing batches.
Why it matches plant phenotyping methods個別種子のRGB画像から種皮色を定量抽出し、発芽予測・事前選別に用いる非破壊的な表現型計測法が研究の中心である。
abstractthe potential of non-destructive, rapid color analysis for predicting germination in coniferous species requires further exploration
Reproduction assets foundThe paper openly deposits its three core phenotyping datasets in Mendeley Data: morphometric seed data (Dataset 1), the raw VIS/RGB scanner images of individual Pinus sylvestris seeds (Dataset 2), and germination outcome data (Dataset 3). All three DOIs are listed in the Data Availability statement and match allowed URDataset · publicThe original morphometric data—Dataset 1—of Pinus sylvestris L. are openly available in Mendeley Data at DOI: https://doi.org/10.17632/8g258nbgmf.1Open asset ↗Mendeley Data · 10.17632/8g258nbgmf.1lines:133-160Dataset · publicThe original VIS image data of Pinus sylvestris L. are openly available in Mendeley Data at DOI: https://doi.org/10.17632/dt78jhyw2j.2Open asset ↗Mendeley Data · 10.17632/dt78jhyw2j.2lines:133-160Dataset · publicThe original germination data—Dataset 3—are openly available in Mendeley Data at DOI : https://doi.org/10.17632/hrs3fgc8tt.1Open asset ↗Mendeley Data · 10.17632/hrs3fgc8tt.1lines:133-160Plant phenotyping relevance match · UnverifiedCrossref · checked 14 Sept 2026
Abstract Monitoring spatial variations in plant growth and forecasting yield before harvest provides valuable insights for optimizing agronomic decision‐making in potato ( Solanum tuberosum L.) cultivation. Although unmanned aerial vehicle (UAV)‐based remote sensing has recently enabled the development of tuber fresh weight (TW) estimation models, their integration into practical yield‐forecasting systems remains limited. In this study, we developed machine learning models to estimate tuber weights at multiple preharvest time points using RGB and multispectral UAV imagery. Image‐derived features were extracted from the orthomosaic and digital surface model images for each plot, and a random forest regression model was trained for TW estimation. The estimated values were subsequently used to fit the Gompertz growth curves, which were then used to forecast the yield at the expected harvest time. The correlation between the estimated and observed values was strong in the UAV‐based TW estimation, with correlation coefficients exceeding 0.8 and coefficients of determination ( R 2 ) above 0.6 at all time points. Yield forecasts based on fitted growth curves achieved a correlation of 0.78 and an R 2 of −0.17 in 2023 and 0.70 and an R 2 of 0.47 in 2024. These results demonstrate that UAV‐based sampling combined with machine learning is a feasible approach for monitoring spatiotemporal variations in tuber growth and forecasting potato yield at the plot level prior to harvest.
Why it matches plant phenotyping methodsUAV画像からジャガイモ塊茎重量を推定し、機械学習と成長曲線で収量を予測する手法が研究の中心であり、推定精度も検証している。
abstractwe developed machine learning models to estimate tuber weights at multiple preharvest time points using RGB and multispectral UAV imagery.
Abstract Digital seed phenotyping offers an efficient and objective alternative to conventional, labor-intensive methods for morphological characterization in plant breeding programs. In castor bean, rapid and reliable tools are essential to support genetic improvement. This study evaluated the potential of digital phenotyping for seed characterization and its application in breeding. Seeds from 65 accessions (2023) and 51 accessions (2024) were photographed with an RGB camera and processed in ImageJ® for extraction of morphological traits. Agreement between digital and manual measurements was assessed by correlation and Bland–Altman analysis, while machine learning models were trained to predict hundred-seed weight (HSW). Genetic diversity was explored using principal component analysis (PCA) and clustering, and variance components and heritability were estimated with mixed linear models. Digital phenotyping showed strong agreement with manual measurements (r = 0.95–0.97) and enabled accurate HSW prediction, with Ridge Regression achieving the best performance (R 2 = 0.88; RMSE = 3.83; MAE = 3.19). PCA explained 85.7% of the variance and revealed three phenotypic clusters. Traits such as seed length (H 2 = 0.88) and aspect ratio (H 2 = 0.87) exhibited high heritability, while roundness (H 2 = 0.79), perimeter (H 2 = 0.72), and area (H 2 = 0.67) were moderate. These findings demonstrate that digital phenotyping is a reliable and high-throughput method for castor bean seed characterization, supporting genotype selection and the integration of machine learning approaches into breeding programs for greater precision and efficiency.
Why it matches plant phenotyping methods種子形態形質の画像取得・抽出、手動測定との技術検証、重量予測モデル評価が研究の中心であり、植物フェノタイピング手法として明確に適格。
abstractDigital seed phenotyping offers an efficient and objective alternative to conventional, labor-intensive methods for morphological characterization in plant breeding programs.
Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · checked 15 Sept 2026
Introduction Reliable identification of plant diseases from leaf images is essential for effective crop monitoring and the prevention of yield deterioration. With the growing adoption of deep learning in agricultural applications, convolutional neural network–based classifiers have demonstrated notable success in visual plant disease recognition. Methods In this study, we propose PlantPathNet , a purpose-built deep learning architecture for plant disease classification. Input images are transformed from the RGB to the HSV color space to enhance the representation of disease-related visual features. A novel Cross-layer Feature Integration Module (CFIM) is introduced to effectively aggregate discriminative features across multiple network depths. Additionally, an efficient channel attention mechanism based on ECANet is incorporated to emphasize disease-relevant representations. The model is optimized using a composite loss function combining modified softmax loss and center loss to address class imbalance and improve feature separability. Results Extensive experiments conducted on the PlantVillage dataset demonstrate that PlantPathNet outperforms several state-of-the-art models, including ResNet-50, Inception-V3, DenseNet121, VGG16, and Vision Transformer-based approaches. The proposed model achieves an overall accuracy of 99.57%, precision of 99.52%, recall of 99.54%, F1-score of 99.53%, and an AUROC of 99.84%. Discussion The results indicate that the integration of HSV-based preprocessing, CFIM, and channel attention significantly enhances classification performance. The proposed framework provides a robust and efficient solution for automated plant disease diagnosis and has strong potential for real-world agricultural applications.
Why it matches plant phenotyping methods葉画像から植物病害状態を推定する深層学習手法を新規開発し、既存モデルとの比較検証も行っているため、植物フェノタイピング手法が中心である。
abstractwe propose PlantPathNet , a purpose-built deep learning architecture for plant disease classification.
RGB / grayscaleFlowerClassificationGrowth / development / phenology
This study presents a comprehensive BDFlower growth stage dataset designed to support research in precision agriculture and floriculture. The dataset encompasses eight common flower species found in Bangladesh: Bush Allamanda, Red Hibiscus, Yellow Bell, Pinwheel Flower, Pink Periwinkle, White Madagascar Periwinkle, Marvel of Peru, and White Hibiscus. Each species is represented across three growth stages-Early, Mid, and Full-resulting in 24 distinct classes. A total of 23,334 colour images are included, comprising 3889 original photographs and 19,445 augmented samples generated with five augmentation techniques. Bush Allamanda contains 499 images, Red Hibiscus contains 489 images, Yellow Bell contains 483 images, Pinwheel Flower contains 497 images, Pink Periwinkle contains 452 images, White Madagascar Periwinkle contains 472 images, Marvel of Peru contains 468 images and White Hibiscus contains 529 images. Each image was collected using smartphone camera at three-time intervals per day, spaced eight hours apart, to capture natural variations in lighting and appearance. The dataset is further organized into training, validation, and testing splits, enabling direct application to machine learning workflows. This is a publicly available dataset specifically curated for flower growth stage classification. In addition to dataset collection, we also conducted a simple experiment using a CNN model to evaluate its performance on this dataset. It is intended to facilitate the development of robust computer vision models that can monitor flower development, with potential applications in automated plant phenotyping, crop monitoring, and digital floriculture systems.
Why it matches plant phenotyping methods花の生育段階を画像で分類する公開データセットを構築し、CNN評価も行っており、植物表現型取得・解析が研究の中心である。
abstractThis is a publicly available dataset specifically curated for flower growth stage classification.
Reproduction assets foundThe paper's own flower growth-stage image dataset (BDFlower) is publicly deposited on Mendeley Data with an explicit direct URL and DOI, directly reproducing the paper's phenotyping (flower growth stage) image measurements. No author analysis code or trained model checkpoints are explicitly deposited.Dataset · publicRepository name: Data Mendeley
Data identification number: 10.17632/m8g2wynwyr.2
Direct URL to data: https://data.mendeley.com/datasets/m8g2wynwyr/2Open asset ↗10.17632/m8g2wynwyr.2html-lines:94-129Plant phenotyping relevance match · UnverifiedOpenAlex · checked 5 Sept 2026
Drought stress poses a significant threat to rose (Rosa spp.) cultivation, impacting plant vigor, floral quality and marketability. Traditional drought screening methods are often destructive and labor-intensive, limiting their application in large-scale breeding programs. This study presents a non-destructive and high-throughput phenotyping approach for assessing drought responses in rose using RGB-derived vegetation indices (VIs) obtained from multi-angle imaging. Twenty-eight diverse rose genotypes were evaluated under well-watered (WW) and induced drought (ID) conditions using a LemnaTec Scanalyzer 3D platform. A total of 56 indices from side-view (SV) and top-view (TV) images were computed to quantify canopy color, greenness and pigment-related traits. Analysis of variance revealed significant genotypic differences and strong genotype × treatment (G × T) interactions across most indices, demonstrating their sensitivity to drought-induced physiological changes. Multivariate analyses, including Principal Component Analysis (PCA) and Pearson correlation matrix evaluation, were performed to explore trait relationships and identify key traits associated with drought stress. These analyses effectively differentiated greenness-related and stress-responsive traits. In addition, the MGIDI analysis integrated all indices and identified ‘Queen Elizabeth’, ‘Jwala’, Rosa chinensis, ‘Sylvia’ and ‘Rose Sherbet’ as the top-performing drought-tolerant genotypes. Integration of leaf wilting scores validated the reliability of these indices as accurate indicators of drought response, with tolerant genotypes exhibiting lower LWS and higher greenness indices. Overall, the study demonstrates that RGB-based high-throughput phenotyping provides a rapid, efficient and scalable method for drought tolerance assessment in roses, offering a valuable tool for accelerating selection in ornamental breeding programs.
Why it matches plant phenotyping methodsRGB画像から植生指数を抽出する高スループット表現型解析手法の開発・実証が研究の中心であり、バラの干ばつ応答を評価する再利用可能なワークフローを提示している。
abstractThis study presents a non-destructive and high-throughput phenotyping approach for assessing drought responses in rose using RGB-derived vegetation indices (VIs) obtained from multi-angle imaging.
Plant diseases have a significant impact on global food security, especially in staple crops like maize (Zea mays). Traditional disease detection systems depend on professional visual inspection, which is labor-intensive, time-consuming, and not scalable for large agricultural areas. Convolutional Neural Networks (CNNs) are used in this study's deep learning (DL) architecture to detect maize leaf diseases accurately and automatically. A curated dataset of approximately 7,000 high-resolution maize leaf photos was created, representing four classes: healthy, Common Rust (Puccinia sorghi), Northern Leaf Blight (Exserohilum turcicum), and Gray Leaf Spot (Cercospora zeae-maydis). Data were sourced from the Plant Village dataset, real-world field collections from Indian farms, and supplemented synthetically to simulate varied climatic circumstances. Advanced methods including as adaptive learning rate scheduling, gradient clipping, and significant data augmentation were used to train a bespoke CNN model that was improved by transfer learning with ResNet50 and VGG16 backbones. The model attained a test accuracy of 98.2%, beating classic machine learning algorithms like SVM (88.5%) and Random Forest (84.3%). Visualization approaches such as feature maps, Grad-CAM, and LIME improved interpretability and showed the model's capacity to locate disease-relevant features. Web-based user engagement is made possible by deployment-ready implementation, which enables farmers to upload leaf photos for immediate diagnosis. With the potential to cut maize crop losses by 20–30%, this research offers a scalable and affordable alternative to early disease detection in precision agriculture. Future research will investigate autonomous farm management with drone-based real-time surveillance and IoT system integration.
Why it matches plant phenotyping methodsトウモロコシ葉の病害状態を画像から推定するCNN手法を開発し、データセット、比較評価、精度検証、解釈性分析まで行っており、植物表現型取得が中心である。
abstractConvolutional Neural Networks (CNNs) are used in this study's deep learning (DL) architecture to detect maize leaf diseases accurately and automatically.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 15 Sept 2026
It has been repeatedly shown that spike productivity is the main component of wheat yield. The main spike parameters related to productivity are size, the number of grains and spikelets per spike, and the presence or absence of awns. In modern genetic research, morphometric analysis of hundreds and thousands of spikes is required to determine the loci that control spike productivity traits. On the other hand, thousands of accessions in modern collections of wheat genetic resources need detailed description. These considerations motivate the development of digital technologies for describing spike traits in wheat, which can be achieved through image analysis methods. These methods allow for automated acquisition of trait values that can serve as the basis for digital plant collections. Here we propose an extended set of spike characteristics obtained both manually and through digital image analysis and present plant characterization. These data form the basis of the updated version of the SpikeDroidDB database (http://spikedroid.biores.cytogen.ru/). The digital description of the spike consists of two blocks. The block of uploaded data includes a description of the plant and contains five tables: collection; variety sample (year of cultivation (vegetation), sowing identifier, taxonomic information, etc.), planting site, and characteristics of the spike determined manually (length, width of frontal and lateral views, type and color of the spike, etc.) The block of extracted features includes spike characteristics obtained by digital phenotyping and contains six tables: characteristics of the spike outline in the image; characteristics of the quadrangle model, values of the color components of the spike, dominant colors of the spike, and texture characteristics of the spike in the image. The most illustrative and significant features of the spike have been identified, allowing for the formation of the spike digital certificate, which includes size, shape, and color features derived from the digital images. The features forming the digital certificate have been compared between two wheat species, T. aethiopicum and T. carthlicum. It is shown that the features of the digital certificate allow for a clear representation of the spike model and the identification of distinct parameters: colors of the spike and awns and roundness of the frontal view of the spike. The database interface has been supplemented with the ability to upload data on plant and spike characteristics, as well as their images, in the batch mode.
Why it matches plant phenotyping methods小麦穂の画像解析による形態形質の自動取得、デジタル表現、データベース基盤を中心に開発・提示しており、植物フェノタイピング手法が研究の中核である。
abstractThese considerations motivate the development of digital technologies for describing spike traits in wheat, which can be achieved through image analysis methods.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 5 Sept 2026
Recent technological advancements employ imaging techniques to examine the morphological, physiological, and genetic differences among plant accessions, enhancing precision and productivity. High-throughput phenotyping serves as an essential method for selecting traits and reducing errors tied to manual data collection. However, the effects of camera-to-object distance in imaging acquisition for plant phenomics have received less attention. We analyzed the imaging parameters that define well the morphological characteristics of pepper and the effects of camera-to-object distance on the imaging of plant growth, leaf dimensions, fruit, and seed characteristics. Three camera-to-object distances (0.8, 1.0, and 1.2 m) were studied for the vegetative stage, and six camera-to-object distances (0.35-0.85 m) were used for the reproductive stages. The results demonstrated that imaging parameters such as Major (the longest line that can be drawn within the leaf) and Minor (the shortest line perpendicular to the major axis) are more effective for assessing canopy spread, while imaging Height provides a strong correlation (r = 0.9) for actual plant height measurement. An optimal camera-to-object distance of 0.8 m yielded better correlations for vegetative traits across all pepper genotypes, likely due to resolution factors at different growth stages. For fruit and seed traits, shorter distances of 0.55 m and 0.65 m were suitable. Additionally, the weights of fresh and dry fruit correlated highly with image area (r = 0.94 and 0.89, respectively, at 0.55 m). The studied pepper genotypes exhibited distinct seed characteristics, including variations in Roundness, Solidity, and Circularity. The imaging approach can accurately capture various plant characteristics and has the potential to replace traditional methods for assessing plants.
Why it matches plant phenotyping methodsRGB画像による植物形質取得を中心に、カメラ距離と撮像パラメータを最適化・検証しており、方法開発および技術検証に該当する。
titleDigitalised phenotyping of pepper (Capsicum spp.) using effective RGB imaging and optimised camera positioning.
Abstract A genome‐wide association study (GWAS) using digital images was conducted to delineate regions of the genome that govern the leaf flipping quantitative trait in soybean ( Glycine max (L.) Merr). However, converting the digital data to numerical scores for downstream analyses was challenging. We have developed an algorithm that operates in the hue, saturation, and value color space in a structured image processing pipeline that includes preprocessing, binary masking for leaf region isolation, contrast enhancement, grid‐based intensity analysis, and thresholding for detecting folded leaves, a response of soybean to drought. The outputs of this image analysis reached over 90% detection accuracy for images captured under different imaging conditions. GWAS using the processed images identified the same genetic loci underlying drought tolerance as were identified earlier by GWAS of the manually curated dataset from the same photos. This approach provides a robust, scalable, and cost‐effective tool for digital image‐based high‐throughput phenotyping.
Why it matches plant phenotyping methods大豆葉の反転表現型を画像から定量化する画像処理アルゴリズムを開発し、異なる撮像条件で精度検証しているため、植物フェノタイピング手法が中心である。
abstractWe have developed an algorithm that operates in the hue, saturation, and value color space in a structured image processing pipeline that includes preprocessing, binary masking for leaf region isolation, contrast enhancement, grid‐based intensity analysis, and thresholding for detecting folded leaves
This dataset consists of a collection of high-resolution RGB images of grapevine leaves, designed to support research in plant pathology, precision viticulture, and computer vision. The images were collected in situ from experimental and commercial vineyards in the north of Portugal, covering different vineyard conditions and management practices. The dataset includes healthy leaves from three grapevine Portuguese cultivars Loureiro, Viosinho and Malvasia Fina, photographed under natural lighting conditions without artificial adjustments. It is organized into five categories: healthy leaves and leaves showing symptoms of downy mildew ( Plasmopara viticola ), powdery mildew ( Erysiphe necator ), Esca complex and Erineum Mite ( Colomerus vitis ). Images are provided in JPEG format with a resolution of 3000 × 3000 pixels and 1024 × 1024 pixels and arranged in folders by health status and disease type. This dataset can be used for machine learning and deep learning applications in disease detection/classification, cultivar identification, and can support other precision agriculture applications, as well as being used for agricultural robotics and educational purposes. An evaluation on three deep learning architectures demonstrated the suitability of the dataset into separating the five classes.
Why it matches plant phenotyping methodsブドウ葉の病徴を画像化した再利用可能なデータセットで、植物の健康状態・病害状態の画像ベース推定を支えることが中心です。深層学習による5クラス分類評価も記載されています。
abstractThis dataset consists of a collection of high-resolution RGB images of grapevine leaves, designed to support research in plant pathology, precision viticulture, and computer vision.
Reproduction assets foundThe paper is a Data in Brief article describing a public Zenodo repository of RGB grapevine leaf images (healthy plus downy mildew, powdery mildew, Esca complex, erineum mite) collected for plant disease/phenotyping research, with explicit data accessibility details. No author analysis code or trained model checkpointsDataset · publicData accessibility
Repository name: Zenodo
Data identification number: https://doi.org/10.5281/zenodo.17343473
Direct URL to data: https://zenodo.org/records/17343473Open asset ↗Zenodo · 10.5281/zenodo.17343473html-lines:93-144Code / dataset availability confirmedOpenAlex · checked 15 Sept 2026
ABSTRACT 3D reconstruction has matured into a robust technology. However, small, flexible objects such as conifer seedlings remain challenging due to their fine‐scale structures, and susceptibility to movement. This study investigates and evaluates methods for reconstructing spruce ( Picea abies ) and pine ( Pinus sylvestris ) seedlings, with the aim of establishing a workflow capable of capturing geometry and texture for applications in machine learning and virtual testing environments. Two acquisition approaches were tested: photogrammetry using a RGB camera and a 3D scanner, both mounted on a robotic arm. While the scanner produced incomplete results, the photogrammetry approach successfully generated point clouds (pcl) with color information. Three different photogrammetry software were tested before relying on Agisoft Metashape and Meshroom for image processing and dense pcl generation, followed by pcl filtering in CloudCompare and meshing in Blender. Six seedlings were reconstructed to textured meshes and quantitatively evaluated using the metrics precision, recall, F1‐score, mask intersection‐over‐union (IoU), and boundary IoU. Results showed an average mask IoU of 75.7% and F1‐score of 86.1%. Pine seedlings yielded higher recall and F1‐scores, whereas spruce reconstructions demonstrated higher precision. The proposed semi‐automated workflow demonstrates the feasibility of reconstructing small and slender structured flexible objects, specifically conifer seedlings.
Why it matches plant phenotyping methods針葉樹苗の形状・テクスチャを取得する3D画像再構成ワークフローを開発・比較・定量評価しており、植物フェノタイピング手法が中心である。
abstractThis study investigates and evaluates methods for reconstructing spruce ( Picea abies ) and pine ( Pinus sylvestris ) seedlings, with the aim of establishing a workflow capable of capturing geometry and texture for applications in machine learning and virtual testing environments.
Reproduction assets foundThe paper's Data Availability Statement states that the raw seedling image data and finalized textured meshes (the paper's phenotyping/3D reconstruction inputs and outputs) are freely available on Zenodo under DOI 10.5281/zenodo.19823955, which appears in the allowed URL list.Dataset · publicand without adjusting the scanning parameters, while also re-
Data Availability Statement
taining texture and color. In contrast to prior approaches that
require manual intervention or do not preserve visual informa- Raw image data and finalized textured meshes are freely available at
Zenodo.org with https://doi.org/10.5281/zenodo.19823955.
tion, the proposed workflow enables a semi-automated recon-
struction process suitable for dataset generation. As shown, the
methodology is effective for the digital reconstruction of small References
and slender structured flexible objects and holds potential for
Abbood, S. A., H. A. Ajjah, A. H. H. Alboabidallah, M. U. MohaOpen asset ↗Zenodopdf-layout-page:14 lines:50-74Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Accurate recognition of potato leaf diseases plays an essential role in precision agriculture and crop protection. Traditional convolutional neural networks achieve promising results but mainly rely on texture and colour cues, neglecting lesion geometry and morphological evolution. This paper introduces a morphology-aware convolution (MAC) framework that embeds learnable morphological operations into deep feature extraction. The learnable morphological kernel adaptively performs differentiable dilation and erosion to capture lesion boundaries, while the structure-guided fusion module integrates structural and textural cues through adaptive weighting. Experiments on the PlantVillage-Potato dataset demonstrate that MAC achieves higher accuracy and robustness than existing convolutional and attention-based architectures with minimal parameter overhead. The method effectively enhances lesion boundary perception and shape consistency, offering improved interpretability and generalisation. These findings suggest that embedding morphological priors into neural networks provides a principled and efficient approach for fine-grained plant disease recognition and other structure-aware visual analysis tasks.
Why it matches plant phenotyping methods植物葉の病徴・病斑形状を画像から認識する新規深層学習手法を開発し、既存手法と精度・頑健性を比較検証しているため、植物フェノタイピング手法が中心です。
abstractThis paper introduces a morphology-aware convolution (MAC) framework that embeds learnable morphological operations into deep feature extraction.
PURPOSE: The leaf area index (LAI) is a crucial parameter for crop growth management. While UAV remote sensing has been utilized to estimate LAI at the plot scale, its application to complex farmland environments—characterized by heterogeneous backgrounds (e.g., soil, residue, and weeds)— has been less explored. METHOD: This study employed UAV-mounted hyperspectral and RGB sensors to gather data from both experimental plots and farmland environments. Data from diverse rapeseed cultivars and growth stages were used as the calibration dataset, while farmland-level data validated the models. The study compared three models: the PROSAIL model, an empirical model incorporating canopy spectral and morphological parameters without differentiating canopy cover types, and the proposed canopy morphological parameters (CMP) model. The CMP model estimated LAI using fractional vegetation cover (FVC) for sparse canopies and canopy height for closed canopies. RESULT: Despite challenges such as UAV image resolution and the limited availability of spatial data, the CMP model showed strong performance, with an R² of 0.779 and RMSE of 0.732. Although its R² was similar to that of the empirical spectral–morphological (ESM) model (R² = 0.780), the CMP approach achieved a notably lower RMSE (0.732 vs. 0.814). This improvement stems from its canopy-aware design, which adaptively uses fractional vegetation cover for sparse canopies and canopy height for closed canopies. Such differentiation enhances model stability and generalization in heterogeneous farmland scenes—conditions in which background interference and structural variability often degrade empirical models. In comparison, the PROSAIL model performed less accurately (R² = 0.618, RMSE = 1.094). CONCLUSION: These results highlight that the CMP model provides a robust and cost-effective solution for LAI estimation, supporting crop growth assessment and management in real farmland.
Why it matches plant phenotyping methodsUAV画像・センサーから rapeseed のLAIを推定するモデルを開発・比較し、異なる圃場条件で検証しているため、植物形質取得手法が研究の中心です。
abstractThe CMP model estimated LAI using fractional vegetation cover (FVC) for sparse canopies and canopy height for closed canopies.
Accurate and non-destructive estimation of rice Leaf Area Index (LAI) is vital for crop growth assessment and yield prediction. Close-range, non-contact optical methods are commonly used for LAI monitoring. However, their accuracy is often limited by the platform, and most rely on single-source data prone to saturation effects and background interference. To overcome these limitations, this study develops a phenotyping robot-based multi-source, high-resolution data fusion approach for field-scale LAI monitoring. A phenotyping robot equipped with multispectral and high-resolution RGB cameras was used to collect vegetation indices, color indices, texture features, and canopy coverage extracted from high-resolution RGB imagery. These features were further combined with meteorological variables to build machine learning models. The Random Forest model achieved the best performance (R² = 0.92, RMSE = 0.302). SHAP (Shapley Additive Explanations) was applied to interpret the model and quantify the importance of multispectral features, RGB-derived texture information, canopy coverage and meteorological factors. Canopy coverage, NDVI and Clgreen were identified as the key factors for improving model performance, and the complementary mechanism between canopy coverage and other features can alleviate the saturation effect in the high LAI stage. The results show that combining high-resolution remote sensing data from robots with meteorological data can effectively mitigate the saturation effect and soil background interference in LAI estimation, and significantly improve the accuracy of LAI estimation. This study provides a practical and scalable framework for field phenotyping and offers technical support for precise rice cultivation and smart agriculture.
Why it matches plant phenotyping methodsロボット搭載マルチセンサーと画像特徴量融合によるイネLAI推定手法を開発・評価しており、表現型取得・推定が研究の中心である。
abstractthis study develops a phenotyping robot-based multi-source, high-resolution data fusion approach for field-scale LAI monitoring.
CONTEXT: The rapid advancement of digital technologies alongside increased global commitment to sustainability has intensified the need for efficient crop management solutions. Across agricultural systems, delayed detection and misclassification of plant diseases remain major contributors to reduced yields, threatening food security and undermining progress toward Sustainable Development Goals such as Zero Hunger, No Poverty, Good Health and Well-being, Climate Action, and Life on Land. Plant pests, diseases, and excessive chemical use further exacerbate these challenges. Early, automated visual detection offers a pathway to environmentally responsible and economically viable agricultural practices. OBJECTIVE: This study aims to develop and evaluate an automated, real-time plant disease classification framework using Vision Transformers (ViT) and hybrid ViT–CNN architectures, with the goal of supporting farmers and agronomists in early decision-making and sustainable crop protection. METHODS: The research employs deep learning techniques, including ViT and a combined ViT–CNN model built on ResNet-9, trained and evaluated using four publicly available datasets: the Turkey Plant Pests and Diseases (TPPD) dataset (15 classes), the Namibia Maize Image Dataset (3 classes), the Banana Image Dataset (3 classes), and the Tanzania Maize Dataset (3 classes). SHapley Additive exPlanations (SHAP) were applied to generate saliency maps for interpretability. Comparative analyses assessed performance, accuracy, and classification speed across attention-based and hybrid architectures. RESULTS AND CONCLUSIONS: The proposed model achieved strong performance with 97.4% accuracy, 96.4% precision, 97.09% recall, a 95.7% F1-score, and high agreement measured by Cohen’s Kappa, outperforming existing benchmark models. SHAP visualizations highlighted that the model leverages high-activation areas, edge features, color patterns, texture, shape, and contextual cues in its predictions. While attention-based models improved accuracy, they also caused reduced classification speed. However, integrating attention blocks with CNN layers effectively compensated for this slowdown, achieving both high accuracy and efficient inference. To evaluate the interpretability and deployment feasibility of the proposed model, we considered several parameters are considered to have an integration of faithfulness, localization quality, sparsity, latency, and energy of the models, including pointing accuracy, localization IoU, Centroid Localization Error (pixels), Attribution Sparsity (%), Insertion AUC, Deletion AUC, Time per Explanation (ms/image), Energy Consumption (J/image), Memory Footprint (MB). SIGNIFICANCE: This research provides a transparent and high-performing deep learning solution for plant disease classification, promoting sustainable agricultural development. By reducing reliance on excessive pesticide and herbicide use, enhancing early diagnosis, and improving decision-making, the model supports environmentally responsible farming practices and contributes to global efforts toward food security and ecological resilience. The significance of this evaluation is that it comprehensively assesses the model’s interpretability and real-world deployability by measuring explanation reliability (faithfulness), spatial precision (localization quality), efficiency (latency and memory), and sustainability (energy consumption), ensuring the model is not only accurate but also transparent, efficient, and practical for deployment.
Why it matches plant phenotyping methods植物病害を画像から直接分類する深層学習フレームワークの開発・評価が研究の中心であり、病害状態という植物表現型を推定するため、対象範囲に含める。
abstractThis study aims to develop and evaluate an automated, real-time plant disease classification framework using Vision Transformers (ViT) and hybrid ViT–CNN architectures
Given the substantial agronomic and economic significance of maize, the development of real-time and high-precision disease detection methodologies is essential for ensuring yield stability. While hyperspectral imaging excels at capturing fine-grained spectral signatures of infection, its higher detection precision comes with considerable high hardware and temporal costs compared to RGB imaging, posing significant challenges for scalable field applications. To bridge this gap, this article proposes a fusion perception unfolding network (FPUF-Net) for high-fidelity maize spectral reconstruction from RGB images. Distinct from conventional deep learning models, FPUF-Net unfolds the optimization problem via a half-quadratic splitting algorithm, solving the data subproblem and prior subproblem alternately during iterations. Specifically, a fusion feature learning network and a spectral-spatial joint attention network are designed within the data subproblem to explicitly exploit RGB spatial priors and mitigate spatial smoothing. Moreover, a spectral-spatial Transformer is utilized as the denoiser to capture long-range spectral dependencies in the prior subproblem. Experiments performed on a maize spectral recovery dataset comprehensively demonstrate that FPUF-Net can effectively reconstruct maize hyperspectral images with superior precision. The structural characteristics enable the network to perceive long-term spectral-spatial fusion features, significantly reducing reconstruction errors, particularly in the biologically critical red-edge region (620-700 nm). In downstream disease detection tasks, the overall accuracy of reconstructed HSIs improves over RGB by margins of 0.51% to 8.1% across different scenarios, while the average accuracy increases by 2.45% to 18.86%. These results indicate that the proposed model offers a viable, cost-effective solution for applying hyperspectral imaging in field settings, enabling its scalable use in agricultural robots.
Why it matches plant phenotyping methodsRGB画像からトウモロコシのハイパースペクトル情報を再構成する手法を開発し、データセットで性能評価している。植物のスペクトル状態および病害検出に直接関わる手法が研究の中心である。
abstractthis article proposes a fusion perception unfolding network (FPUF-Net) for high-fidelity maize spectral reconstruction from RGB images.
PURPOSE: Above-ground biomass (AGB) is a critical indicator for assessing the growth status of winter wheat. Since the range of extracted color indices (CIs) tends to remain constant after flowering and UAV RGB images cannot capture the lower and middle structures of the canopy under dense planting conditions, the efficiency of AGB estimation models is limited. Therefore, this study aimed to improve the accuracy of winter wheat AGB estimation by incorporating canopy volume information with RGB-based CIs. METHODS: RGB images were acquired to generate Digital Orthophoto Maps (DOM) and Digital Surface Models (DSM) at Feekes 10, Feekes 10.5.2, Feekes 10.5.4, and Feekes 11.3 growth stages. Eight biomass-related CIs were extracted from the DOM, and canopy volume (V) was calculated from the DSM for corresponding regions. The RReliefF algorithm was applied to rank feature importance and select optimal features. Eight statistical and machine learning regression algorithms, including Gaussian process regression (GPR), were used to construct AGB estimation models with different feature combinations. RESULTS: The results showed that the GPR algorithm outperformed other regression methods, achieving the highest estimation accuracy with R² values of 0.775, 0.741, 0.702, and 0.568 at the four growth stages, respectively. Compared with models using CIs alone, integrating canopy volume with CIs improved AGB estimation accuracy from Feekes 10.5.2 to Feekes 11.3, with R² increases of 7.31%, 6.55%, and 22.98%, respectively. CONCLUSION: Overall, combining canopy volume features derived from UAV-based RGB imagery with CIs and applying effective machine learning algorithms enables rapid and accurate estimation of winter wheat AGB.
Why it matches plant phenotyping methodsUAV画像から色指数とキャノピー体積を抽出し、冬コムギの地上部バイオマスを推定する取得・解析手法が研究の中心であるため。
titleWheat biomass estimation by fusing color index and canopy volume based on UAV RGB images
• The investigation of the phenotypic variation of F. vesca has provided significant insights into the factors influencing fruit traits, including shape, size, and overall quality. • Digital phenotyping has been proven to be more effective in distinguishing F. vesca fruit traits that are difficult to phenotype in the selection process. • The integration of digital phenotyping and genomic analysis have proven to be a powerful strategy for improving the selection of quality traits in F. vesca . • Genome-based breeding and predictive models facilitate the development of climate- resistant F. vesca cultivars that meet consumer preferences. The quality of Fragaria vesca berry fruits is an important factor in their marketability, and therefore, it has become a major target of breeding programs. However, berry traits are difficult to dissect due to the complex interaction of genetic and environmental factors. In this study, we evaluated phenotypic variation in commercially relevant traits, including shape, size, pH and total soluble solids (SSC) in an open-pollinated F. vesca population grown in a region characterized by high temperature fluctuations. The observed variability underscores the intricate interplay between genetic background and environmental factors on fruit morphology and quality traits. Digital imaging phenotyping proved to be a robust and objective approach for capturing morphological traits difficult to phenotype, providing quantitative data necessary for effective selection. Moreover, the ddRAD sequencing facilitated the identification of significant genetic diversity in a F. vesca population, generating approximately 4000 SNP polymorphic markers used to investigate the population structure and the potential of genomic models for selecting desirable traits, such as fruit shape and size. Several genomic selection models were tested to predict breeding values for fruit morphological traits. Prediction accuracy was substantially improved through training set optimization strategies, particularly those based on CDmean criteria. The integration of digital phenotyping, high-throughput genotyping and genomic predictive modelling has proven to be a powerful strategy for improving the selection of desirable traits in F. vesca . Overall, the findings from this study provide a foundation for further genetic improvement efforts, which will ultimately enhance the quality and marketability of strawberry cultivars.
Why it matches plant phenotyping methodsイチゴ果実の形状・サイズを対象に、デジタル画像による表現型取得を明示的に評価しており、育種への応用だけでなく形態形質の定量的取得が研究の中心的要素である。
abstractDigital imaging phenotyping proved to be a robust and objective approach for capturing morphological traits difficult to phenotype, providing quantitative data necessary for effective selection.
Agriculture is a fundamental component of human civilization. It contributes to the economy while also providing sustenance. Plant foliage or crops are susceptible to many illnesses during agricultural agriculture. The illnesses impede the development of their respective species. Timely and accurate identification and categorization of illnesses may mitigate the risk of further harm to the plants. The identification and categorization of these disorders have emerged as significant challenges. The conventional methods used by farmers to anticipate and categorize plant leaf diseases may be tedious and inaccurate. Challenges may occur while endeavouring to manually forecast illness kinds. The failure to promptly identify and categorize plant diseases may lead to the devastation of crops, causing a substantial reduction in yield. Agriculturalists using computerized image processing techniques in their fields may mitigate losses and enhance output. A multitude of strategies has been used in the identification and categorization of plant diseases using photographs of sick leaves or crops. In this research, convolutional neural networks (CNNs) are often used for image recognition and classification because of their intrinsic ability to autonomously extract relevant visual characteristics and comprehend spatial hierarchies. Consequently, in many sophisticated image recognition and classification tasks, deep learning, mostly via convolutional neural networks, is favoured when substantial data and computing resources are accessible, demonstrating effective detection and classification outcomes on their datasets. This methodology seeks to enhance productivity, minimize crop losses, and foster sustainable agricultural practices via the provision of valuable information and the automation of disease identification. The multilingual solution guarantees inclusion for diverse agricultural communities by automating disease detection and providing actionable information.
Why it matches plant phenotyping methods病葉画像から植物病害を検出・分類する画像解析手法が研究の中心であり、植物の病害状態を直接推定しているため、植物フェノタイピング手法として採用する。
abstractA multitude of strategies has been used in the identification and categorization of plant diseases using photographs of sick leaves or crops.
Abstract Sugarcane ( Saccharum spp.), a C4 plant, is a vital renewable biofuel and sugar source for industries worldwide. However, synchronizing flowering between parental lines often poses challenges for breeders, hindering effective crossbreeding efforts. This study aimed to develop a high‐throughput phenotyping (HTP) strategy to evaluate flowering‐related traits using vegetation indices (VIs) and other metrics alongside artificial intelligence (AI)‐based prediction methods. A total of 154 genotypes were planted in an augmented block design at the IAC sugarcane breeding station in Serra Grande‐BA, Brazil. Raw RGB (Red, Green, Blue) images were captured using a DJI Mavic 3 Enterprise drone during the plant cane (PC) and first ratoon (FR) crop seasons. These images were processed to create orthomosaics and compute metrics/vegetation index; subsequently, machine learning (ML) and deep learning pipelines for systematic analysis were developed. A convolutional neural network (CNN) model achieved promising results, with an accuracy rate of up to 84% in the flowering detection task. Additionally, flower counts from the CNN model showed a moderate correlation with field data, evidenced by an R 2 value of 0.72 at the onset and an R 2 value of 0.29 at the conclusion of the flowering season for the PC. This resulted in an overall average regression R 2 of 0.46 with a root mean square error (RMSE) of 13.80. Furthermore, an artificial neural network classification model reached a notable accuracy of 0.87 in differentiating genotypes based on their flowering response (early‐flowering vs. late‐flowering), utilizing VIs and digital model‐based metrics as input parameters. The ML regression model demonstrated performance levels of R 2 = 0.51 and RMSE = 8.06 for days to flag leaf emergence in PC and R 2 = 0.52 and RMSE = 7.93 for days to flowering in FR. These results highlight the potential of HTP strategies, utilizing orthomosaics and AI, to accelerate data collection and analysis, offering significant insights for breeding programs in sugarcane.
Why it matches plant phenotyping methodsドローンRGB画像、オルソモザイク、植生指数、AIを組み合わせた開花形質のハイスループット取得・予測手法を開発し、精度検証まで実施しており、フェノタイピング手法が研究の中心である。
abstractThis study aimed to develop a high‐throughput phenotyping (HTP) strategy to evaluate flowering‐related traits using vegetation indices (VIs) and other metrics alongside artificial intelligence (AI)‐based prediction methods.
Plant breeding has been used for over 10,000 years to adapt crops like wheat to different environments. Continuous breeding efforts have led to high-yielding and resilient wheat varieties, with Canada being the 6th largest producer in 2024-25. However, challenges like leaf rust, caused by Puccinia triticina Erikss., still impact global wheat production. Rapid pathogen evolution requires ongoing identification and deployment of novel resistance sources. Accurate phenotyping of key traits like disease resistance, winter survival, and plant height is a major bottleneck in wheat breeding. Advances in high-throughput (HTP) genotyping and phenotyping offer new opportunities to enhance genetic gain. This study integrated genetic mapping and UAV-based HTP approaches to evaluate leaf rust resistance and key agronomic traits in winter wheat. A doubled-haploid (DH) population (n = 130) developed from the cross W538/Emerson was evaluated for leaf rust resistance at the seedling and adult plant stages. Genotyping was performed using a 25K Infinium SNP array, and linkage and QTL analyses mapped resistance genes. Seedling-stage resistance was associated with a locus on chromosome 1B, while adult plant resistance was governed by multiple QTL, including QLr.umb-1B, QLr.umb-2A, QLr.umb-3B, and QLr.umb-4D. Lines carrying multiple resistance QTL exhibited enhanced leaf rust resistance, highlighting the importance of QTL stacking for durable resistance. UAV-based HTP methods were evaluated for assessing spring stand and plant height in winter wheat breeding nurseries. Manual ratings were compared with RGB and multispectral-based UAV metrics, including relative plant pixel area and vegetation indices like NDVI and EPVI. NDVI was the most robust method for spring stand assessment, with four times higher heritability than manual ratings. Manual measurements were more accurate than UAV-based methods for plant height, but SfM and LiDAR had comparable performance. This study highlights the complementary value of genetic mapping and UAV-based HTP in wheat breeding, emphasizing multi-QTL resistance for leaf rust and the potential of UAVs to improve phenotyping efficiency for key agronomic traits.
Why it matches plant phenotyping methodsUAV画像・マルチスペクトル・SfM・LiDARによる春季スタンドと草丈の測定を、手動評価と比較・検証しており、植物表現型取得法が研究の実質的な構成要素である。
abstractUAV-based HTP methods were evaluated for assessing spring stand and plant height in winter wheat breeding nurseries.
Early and accurate detection of pigeon pea leaf diseases is essential for improving crop productivity and ensuring food security, particularly under real-field agricultural conditions. This paper introduces a shallow and computationally off-the-shelf deep learning system to detect the presence of pigeon pea leaf disease with great accuracy and in real-time on resource-limited cameras. DSLR and smartphone cameras were used to make up a custom high-resolution dataset under natural field conditions, including healthy leaves and major diseases, such as Fusarium wilt, leaf spot, and powdery mildew. All the images were downsampled to 224 × 224 pixels and processed with a Gaussian smoothing filter to remove noise and a Canny edge detector to improve structural features. Disease regions were accurately isolated using a Skill Optimization Algorithm (SOA)-driven segmentation strategy that dynamically optimized threshold levels, morphological kernel sizes, and lesion area constraints to handle background clutter and illumination variations. A pretrained EfficientNet-B0 model was used to extract deep semantic features, which consisted of compact 1280-dimensional feature vectors. A novel FMDDCN approach was used to classify these features through exploiting the sensitivity to subtle disease patterns by relying on differential feature modeling and multi-layer fusion of features. The model was fitted on stochastic gradient descent with a learning rate of 1 x 10-3 and a batch size of 32, and assessed on a 60/20/20 train validation test split with 5-fold cross-validation. The results of the experiment show consistent convergence with low overfitting. The proposed framework was found to produce a classification accuracy of 94.5%, precision of 91.0%, recall of 85.5% and Matthews Correlation Coefficient of 88.5% when it was used with four optimized features. In comparison, it is demonstrated that FMDDCN performs better than traditional machine learning and deep learning models, with its F1-score of 0.965 and the overall accuracy of 0.965. The suitability of the real-time edge deployment is verified, as confirmed by the use of computational analysis to reduce inference latency and memory consumption.
Why it matches plant phenotyping methods画像から植物葉の病徴・病害状態を推定するリアルタイム手法の開発と技術評価が中心であり、植物フェノタイピング手法に該当する。
abstractThis paper introduces a shallow and computationally off-the-shelf deep learning system to detect the presence of pigeon pea leaf disease with great accuracy and in real-time on resource-limited cameras.
Abstract Abstract. Accurate and timely identi cation of plant diseases is essential for improving crop productivity and ensuring sustainable agricultural practices. This paper presents a comprehensive image dataset of fruit and leaf diseases covering six economically important horticultural crops: Apple, Banana, Citrus, Guava, Mango, and Papaya. The dataset comprises high-quality RGB images representing both healthy and diseased samples, with disease symptoms including spots, lesions, discoloration, blight, rot, and fungal and bacterial infections captured under diverse real-world conditions. Variations in illumination, background complexity, viewing angles, growth stages, and symptom severity are intentionally included to enhance the robustness and generalizability of learning models developed using this data. The dataset is structured in a class-wise manner and preprocessed to support direct integration with deep learning frameworks. It is extensively used to train, validate, and evaluate deep learning based plant disease classi cation models, enabling automatic feature learning from raw images without manual intervention. Experimental usage demonstrates that the dataset is well suited for convolutional neural networks and attentionbased architectures, facilitating e ective discrimination between multiple disease categories across di erent crops and plant organs. By providing a uni ed multi-crop, multi-disease benchmark, this dataset aims to accelerate research in automated crop disease diagnosis, precision agriculture, and intelligent decision-support systems for sustainable farming.
Why it matches plant phenotyping methods植物の葉・果実の病徴画像を収録したデータセット／ベンチマークであり、病害状態の画像ベース推定を中心的に扱うため。
abstractThis paper presents a comprehensive image dataset of fruit and leaf diseases covering six economically important horticultural crops
Reproduction assets foundThe paper's core asset is the ABCGMP fruit and leaf disease image dataset, publicly deposited on Mendeley Data, with author analysis code also stated to be available on GitHub. Both are paper-specific, public, and actionable.Dataset · publicData is available on Mendeley:1Open asset ↗pdf-page:33 lines:1-56Code · publicCode availability: Code is available on GitHub 2Open asset ↗GitHubpdf-page:33 lines:1-56Plant phenotyping relevance match · UnverifiedOpenAlex · checked 15 Sept 2026
Abstract Image processing techniques for plant phenotypes are rapidly evolving, allowing faster, non-destructive and objective evaluation of plant growth parameters than conventional methods for plant physiological studies. The aim of this study was to evaluate the green area rate (GAR) per plant through smartphone images of magnolia, and to predict the differences in the growth parameters including root length from the differences in green area rate of ordinary saplings and ones grown using cerium chloride (CeCl3). An image analysis program developed using fuzzy C-means clustering algorithm from digital images was used to estimate GAR of magnolia saplings, and the correlation between green area rate and growth parameters was determined and regression model was constructed. Correlation coefficient of GAR was the highest with leaf length of 0.98 and lowest with above ground total dry mass of 0.87. Accuracy between the predicted values and measured values was estimated, and then difference of GAR between control saplings and ones grown using cerium chloride was calculated. Based on the changes in green area percentage, we predicted difference in growth parameters. No significant difference existed between the predicted and the measured values. The results obtained in present paper will play a significant role in predicting the growth parameters of saplings without affecting the growth and in detecting effectiveness of various growth promoters.
Why it matches plant phenotyping methodsスマートフォン画像とファジーC-meansによる緑色面積率推定を開発し、植物の成長形質を予測・検証しており、表現型取得・抽出法が研究の中心である。
abstractAn image analysis program developed using fuzzy C-means clustering algorithm from digital images was used to estimate GAR of magnolia saplings
Accurate fruit detection in citrus orchards is essential for yield estimation, precision harvesting, and automated orchard monitoring. Although UAV-based imaging has become a powerful tool in precision agriculture, publicly available datasets for orange fruit detection remain scarce, particularly those integrating multispectral data under real field conditions. This lack of open resources limits the development and benchmarking of robust deep-learning models for cross-spectral and illumination-invariant detection. We present CampanetaOrangeFruit, a dataset acquired with a DJI Mavic 3 Multispectral UAV flying at 14 m above ground level over a commercial citrus orchard in Corbera, Valencia, Spain. The dataset comprises 550 synchronized captures (RGB + four multispectral bands: R, G, RE, NIR) for a total of 2750 images and 301,232 annotated orange instances. Each image includes YOLOv5-format annotations generated through a homography-based reprojection process, ensuring geometric consistency across spectral modalities. CampanetaOrangeFruit uniquely provides pixel-aligned, cross-spectral UAV imagery with fine-grained fruit-level annotations, enabling research on fruit detection, yield estimation, and domain adaptation in real-world orchard environments. It represents a valuable benchmark for advancing deep-learning approaches in precision agriculture and sustainable citrus production.
Why it matches plant phenotyping methods柑橘果実を対象としたUAV画像データセットとアノテーションを提供し、果実検出・収量推定モデルの開発およびベンチマークを中心課題とするため、植物フェノタイピング用データセットとして採用する。
abstractpublicly available datasets for orange fruit detection remain scarce
This study addresses the challenge of detecting white grape clusters (Vitis vinifera L) in high-density vineyard canopies, a critical task for precision viticulture and yield estimation. Traditional statistical and image-processing methods have struggled with occlusion issues. In this work, over 100 field RGB images were collected at La Bergonza (Toledo, Spain) and expanded through data augmentation, with various preprocessing strategies tested to enhance cluster visibility. Convolutional Neural Network (CNN) architectures were compared, highlighting YOLOv8 as superior to Mask R-CNN in both accuracy and efficiency. YOLOv8, trained for up to 100 epochs on equalized and augmented datasets, achieved outstanding performance: 84.9% precision, 72.6% recall, and mAP@0.5 of 83%, far surpassing Mask R-CNN (17% precision, 26% recall). The model successfully detected partially hidden clusters, including those invisible to human experts, better than previous studies that required controlled backgrounds or artificial lighting. Results confirm that combining RGB equalization with data augmentation optimizes detection. These findings underscore the potential of deep learning and low-cost RGB imaging systems to enable automated, scalable solutions for yield estimation and canopy analysis. In conclusion, YOLOv8 emerges as a promising tool for accurate grape bunch detection under field conditions, overcoming previous limitations.
Why it matches plant phenotyping methodsブドウ房を対象としたRGB画像とCNNによる検出手法を開発・比較し、精度を定量評価しているため、植物器官の表現型取得が中心である。
abstractIn this work, over 100 field RGB images were collected at La Bergonza (Toledo, Spain) and expanded through data augmentation, with various preprocessing strategies tested to enhance cluster visibility.
Reproduction assets foundThe paper's Data Availability Statement points to the authors' public GitHub repository containing the original grape-cluster image dataset and annotations used in this study. The ultralytics repository is a generic third-party library, not a paper-specific asset.Dataset · publicData Availability Statement: The original data presented in the study are openly available at
[https://github.com/upmValeriano/racimosUva.git.]Open asset ↗upmValeriano/racimosUvapdf-page:13 lines:1-66Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · checked 5 Sept 2026
Seed varietal purity and physiological viability are critical determinants of crop yield and quality. However, non-destructive assessment faces significant challenges in fine-grained variety discrimination and the perception of internal defects. This study proposes S3-Net, an AI-driven multimodal sensing framework that integrates vision–language alignment with dual-spectral sensor fusion for autonomous seed quality evaluation. We introduce a Knowledge–Vision Alignment (KVA) module that incorporates encyclopedic morphological descriptions to guide feature learning, significantly enhancing few-shot generalization. Complementarily, a Dual-Spectral Fusion (DSF) module combines high-resolution RGB textures with penetrative Short-Wave Infrared (SWIR) sensing to jointly characterize external and internal traits. Experimental results on a custom multimodal dataset of 6000 samples across 12 crop categories demonstrate that S3-Net achieves 96.9% accuracy for species identification and 95.8% for viability detection. Notably, S3-Net outperforms ResNet-50 by 40.3% in extreme 1-shot scenarios. With a stable inference throughput of 95 fps, the system meets the high-throughput demands of industrial-scale applications, providing a robust and efficient solution for intelligent agricultural phenotyping.
Why it matches plant phenotyping methods種子の生理的状態（viability）をRGB・SWIR融合で非破壊推定するセンシング／AI手法が研究の中心であり、データセットと性能評価も提示しているため。種識別のみなら対象外だが、viability検出は植物状態のフェノタイピングに該当する。
abstractThis study proposes S3-Net, an AI-driven multimodal sensing framework that integrates vision–language alignment with dual-spectral sensor fusion for autonomous seed quality evaluation.
Plant diseases pose a serious threat to global food security by directly affecting agricultural production. Traditional expert observation-based diagnosis processes are time-consuming, subjective, and error-prone, making early and accurate diagnosis difficult. This has necessitated the development of image processing and artificial intelligence-based systems that can automatically recognize disease symptoms from leaf images. This study aims to automatically classify plant leaf diseases using the PlantDoc dataset, which consists of images collected under real-world conditions. First, various image processing steps, such as denoising, color space transformations, segmentation, and contour detection, were applied to the leaf images to extract color, texture, and geometry-based features. The resulting features were classified using Support Vector Machines, Random Forests, and k-Nearest Neighbors, and the performance of these models was compared. Furthermore, a deep learning-based MobileNetV2 model was trained using transfer learning and data augmentation techniques and compared with classical methods. Experimental results show that the Random Forests model achieved the highest accuracy rate among classical methods, at 81.5%, while the MobileNetV2 model outperformed all other methods, with an accuracy rate of 86.9%. These findings demonstrate that deep learning-based approaches have higher generalization capabilities on complex, multi-class real-world data. Furthermore, classical methods, thanks to their interpretability and low computational cost, can be a good alternative in resource-limited systems.
Why it matches plant phenotyping methods植物葉の病徴を画像から抽出・分類する画像処理および機械学習手法が研究の中心であり、植物病害状態の表現型推定に該当する。
abstractThis study aims to automatically classify plant leaf diseases using the PlantDoc dataset, which consists of images collected under real-world conditions.
Climate change-driven bark beetle outbreaks pose severe threats to Norway spruce forests across Europe, yet early detection of infested trees remains difficult with conventional field surveys or satellite imagery. Existing detection methods, such as traditional and deep learning-based methods, require extensive time, training data, and computational resources while underutilizing dense temporal observations. We present an integrated UAV framework that merges deep learning-based tree detection with multitemporal spectral analysis for early, tree-level identification of bark beetle stress. A key novelty of this research is the use of high-frequency UAV time series to enable temporal change-detection methods such as CUSUM, providing scalable early-warning capabilities previously unattainable at individual-tree resolution. Firstly, a novel Fused YOLO-SAM pipeline was developed to automatically detect and delineate Norway spruce trees using UAV imagery collected over the mixed forests. Three YOLO v12 models were trained separately on summer (June), fall (November), and combined seasonal datasets acquired with a DJI Phantom drone having 5-band camera. The fall-trained model achieved the highest detection performance (F1 = 0.827, mAP50 = 0.916) and demonstrated superior cross-seasonal generalization, producing accurate and transferable crown segmentations when integrated with the Segment Anything Model (SAM). These automatically delineated spruce canopies provided a consistent and scalable base for individual tree-level spectral analysis. To discriminate between healthy and bark beetle-infested Norway spruce trees at the individual tree level, the delineated crowns were subsequently used to extract per-tree spectral time series for stress analysis. Radiometrically calibrated multispectral bands and vegetation indices were evaluated using cumulative sum (CUSUM)-based Receiver Operating Characteristic (ROC)-Area under Curve (AUC) and effect size metrics. The results demonstrated that calibration significantly improves early-season discrimination, with near-infrared and red-edge bands achieving stable performance (mean AUC ≈ 0.75-0.80; Cohen’s d ≈ 0.9-1.2). In contrast, vegetation indices such as NDRE and MSR-RE exhibited lower discrimination and greater temporal variability. This research also performed comparisons of calibrated multispectral features with uncalibrated RGB features. The analysis indicated that selected RGB indices (notably ExG and GCC) is comparable to the performance of the multispectral bands (REG, NIR), showing positive early discrimination and mean AUC values around 0.70-0.75 which can provide operational benefit for forest monitoring and management when the expensive multispectral cameras are unavailable. Overall, this research demonstrates a scalable and operationally feasible pipeline that integrates automated tree delineation with robust spectral stress diagnostics, supporting early bark beetle detection and informed forest management under resource and data constraints.
Why it matches plant phenotyping methodsUAV画像による樹冠抽出と時系列スペクトル解析を統合し、個体レベルのトウヒの樹皮甲虫ストレスを推定する手法が研究の中心であるため。
abstractWe present an integrated UAV framework that merges deep learning-based tree detection with multitemporal spectral analysis for early, tree-level identification of bark beetle stress.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 5 Sept 2026
max L. Merr.) pubescence color is a trait commonly recorded by breeding programs. In previous research using high-throughput phenotyping (HTP), researchers could separate gray pubescence from light tawny and tawny pubescence, but could not separate light tawny from tawny. Using the Random Forest algorithm and time series of aerial RGB (red, green, blue) and multispectral images, this study aimed to classify pubescence color by testing models in data subsets from experiments grown over three years. By incorporating the pubescence color of the parental lines and training the model with a time series of images (four drone flights before or at maturity), a higher overall accuracy was achieved compared to a single flight at maturity. The red/blue index was the most successful feature for discriminating pubescence color, and the blue normalized difference vegetation index (NDVI) and green NDVI were also helpful, mainly in discriminating gray from light tawny pubescence. The overall accuracy was 86.55% in the best scenario (Kappa = 0.7976), and the sensitivity for gray, light tawny, and tawny pubescence were 0.893, 0.788, and 0.915, respectively. When models were tested in an independent environment, they achieved a lower overall accuracy of 65.86%, but still demonstrated fair to good model reliability (Kappa = 0.4874). Applying an HTP pipeline, as used in this study, would help breeding programs save time classifying pubescence color. Since pod color interferes with this trait in the background, genotyping a proportion of the plant rows for both traits and phenotyping pod color could improve the results.
Why it matches plant phenotyping methods航空画像とRandom Forestを用いてダイズの毛茸色という植物形質を分類し、時系列画像、特徴量、独立環境での精度を検証したHTP手法研究であり、表現型取得・抽出法が中心である。
abstractUsing the Random Forest algorithm and time series of aerial RGB (red, green, blue) and multispectral images, this study aimed to classify pubescence color by testing models in data subsets from experiments grown over three years.
Severe foliage occlusion and dynamically changing lighting conditions in complex orchard environments pose significant challenges for visual perception systems in automated apple harvesting, including low detection accuracy, poor robustness, and insufficient real-time performance. To address these issues, this study proposes an improved lightweight detection network based on YOLOv11, named YOLO-WBL, along with a precise yield estimation algorithm based on 3D point clouds, termed CLV. The YOLO-WBL network is optimized in three aspects: (1) A C3K2_WT module integrating wavelet transform is introduced into the backbone network to enhance multi-scale feature extraction capability; (2) A weighted bidirectional feature pyramid network (BiFPN) is adopted in the neck network to improve the efficiency of multi-scale feature fusion; (3) A lightweight shared convolution separated batch normalization detection head (Detect-SCGN) is designed to significantly reduce the parameter count while maintaining accuracy. Based on this detection model, the CLV algorithm deeply integrates depth camera point cloud information through 3D coordinate mapping, irregular point cloud reconstruction, and convex hull volume calculation to achieve accurate estimation of individual fruit volume and total yield. Experimental results demonstrate that: (1) The YOLO-WBL model achieves a precision of 93.8%, recall of 79.3%, and mean average precision (mAP@0.5) of 87.2% on the apple test set; (2) The model size is only 3.72 MB, a reduction of 28.87% compared to the baseline model; (3) When deployed on an NVIDIA Jetson Xavier NX edge device, its inference speed reaches 8.7 FPS, meeting real-time requirements; (4) In scenarios with an occlusion rate below 40%, the mean absolute percentage error (MAPE) of yield estimation can be controlled within 8%. Experimental validation was conducted using apple images selected from the dataset under varying lighting intensities and fruit occlusion conditions. The results demonstrate that the CLV algorithm significantly outperforms traditional average-weight-based estimation methods. This study provides an efficient, accurate, and deployable visual solution for intelligent apple harvesting and yield estimation in complex orchard environments, offering practical reference value for advancing smart orchard production.
Why it matches plant phenotyping methodsリンゴ果実の検出と3D点群による個別果実体積・総収量推定を開発し、精度・速度・遮蔽条件下で検証しており、植物形質取得が中心的な方法論的貢献である。
abstractthis study proposes an improved lightweight detection network based on YOLOv11, named YOLO-WBL, along with a precise yield estimation algorithm based on 3D point clouds, termed CLV.
Sugar beet is a major sugar crop in temperate regions and rapid, high-throughput, and accurate estimation of field phenotypes is essential for variety selection and production optimization. In this paper, ten commercial sugar beet varieties adapted to high latitudes are investigated using unmanned aerial vehicle (UAV) based red-green-blue (RGB), multispectral, and thermal infrared imaging across multiple growth stages. Canopy structural, texture, spectral, and temperature features are extracted, and three machine learning algorithms, random forest (RF), partial least squares (PLS), and support vector machine (SVM), are used to predict sugar content, root fresh weight, and yield. The results show that all three methods estimate sugar content well, with relative root mean square error (rRMSE) values below 11.0%, while RF and PLS outperform SVM. Multispectral features provide higher accuracy than RGB features, and multi-sensor feature combinations generally improve sugar content prediction compared with single-sensor inputs. For root fresh weight, SVM slightly outperforms RF and PLS, and RGB features are more informative than multispectral features. The integration of thermal infrared features does not notably improve RF or PLS models, but the combination of multispectral and thermal infrared features achieves the best SVM performance ( R2=0.58, RMSE = 75.3 g, and rRMSE = 23.7%). For yield estimation, RF achieves the highest accuracy, with rRMSE values ranging from 15.4% to 18.8%. Yield prediction accuracy increases as the time of image acquisition approaches harvest, and combining multi-temporal data from periods close to harvest further improves model performance. Overall, multi-sensor UAV data can effectively estimate sugar content, root fresh weight, and yield in sugar beet, providing a useful approach for phenotypic analysis, precision management, and variety selection.
Why it matches plant phenotyping methodsUAVマルチセンサー画像から糖含量、根 fresh weight、収量という植物形質を抽出・推定し、センサー特徴量と機械学習モデルの性能を比較しているため、表現型取得・推定手法が中心である。
abstractrapid, high-throughput, and accurate estimation of field phenotypes is essential for variety selection and production optimization.
ABSTRACT Potato common scab ( Streptomyces sp.) is an economically important disease that reduces the quality and market value of tubers. A key aspect in developing management strategies involves accurately quantifying the disease. Due to the three-dimensional nature of the tuber and the heterogeneous distribution of lesions across its surface, visual estimates of severity can be challenging. Therefore, the objectives of this study were to develop and validate a standard area diagram (SAD) for estimating common scab severity on potato tubers and to compare validation outcomes obtained using real tubers and digital images. A SAD comprising six severity levels (from 1.3 to 66.8%) was developed based on image analysis of naturally infected tubers. Validation was conducted using two complementary approaches in which inexperienced raters evaluated either real potato tubers or digital images of the same tubers under unaided and aided conditions. Accuracy, bias components, and inter-rater reliability were quantified using absolute error metrics, Lin’s concordance correlation coefficient, intraclass correlation coefficients, and overall concordance correlation coefficients. Use of the SAD significantly improved accuracy, reduced systematic bias, and increased inter-rater reliability across both validation approaches. No significant differences were detected between assessments conducted on real tubers and images, although image-based evaluations showed a slight, non-significant tendency toward reduced scale and location bias under aided conditions. These results demonstrate that a dimension-aware SAD integrating information across the full tuber surface enhances the reliability and reproducibility of visual severity assessments and supports the use of image-based evaluations for training, large-scale surveys, and remote or collaborative applications involving three-dimensional plant organs.
Why it matches plant phenotyping methodsジャガイモ塊茎の病徴重症度を定量化する標準面積図を開発・検証し、画像評価と実物評価の性能を比較しているため、植物フェノタイピング手法が研究の中心である。
abstractThese results demonstrate that a dimension-aware SAD integrating information across the full tuber surface enhances the reliability and reproducibility of visual severity assessments and supports the use of image-based evaluations
Existing tomato datasets often focus on short-term experiments or lack integrated environmental and agronomic data. We present Horti-M3-Tomato, a comprehensive three-year dataset collected in Northeast China's greenhouse, including high-resolution RGB images, environmental sensor data (recorded every 30 minutes), soil conditions, and detailed agronomic records such as yield data and management practices. Spanning three growing seasons (2023-2025), the dataset integrates temporal imaging, environmental monitoring, soil data, and manual phenotypic and yield records. Horti-M3-Tomato supports research on growth dynamics, genotype-environment interactions, and provides a benchmark for AI-based phenotyping and precision horticulture. The dataset is openly available for further research in controlled-environment agriculture.
Why it matches plant phenotyping methodsトマトの画像・環境センサーデータ・手動表現型記録を統合したデータセットであり、AIベースの表現型解析のベンチマークとして明示されているため、表現型データ基盤が中心です。
abstractincluding high-resolution RGB images, environmental sensor data (recorded every 30 minutes), soil conditions, and detailed agronomic records such as yield data and management practices.
Accurate and non-destructive evaluation of grape quality is crucial for intelligent viticulture, yet most existing approaches address cultivar classification and soluble solid content (SSC) prediction as independent tasks based on single-modality data, limiting robustness and practical applicability. This study proposes DualStream-RTNet, a unified multimodal deep learning framework that simultaneously performs grape cultivar classification and SSC prediction by integrating RGB-HSV fused images and PCA-compressed hyperspectral spectra. The dual-stream architecture enables the complementary learning of external chromatic-textural cues and internal physicochemical information, while a Transformer-enhanced fusion module strengthens global representation and cross-modal correlation. A dataset of 864 berries from five grape cultivars was used to validate the model. DualStream-RTNet achieved 93.64% classification accuracy, outperforming ResNet18 and other CNN baselines, and produced more compact and consistent confusion-matrix patterns. For SSC prediction, it consistently yielded the highest performance across cultivars, with R2p values up to 0.9693 and RMSE as low as 0.2567, surpassing the PLSR, SVR, LSTM, and Transformer regression models. These results demonstrate the superiority of the proposed framework in capturing both visual and spectral characteristics. DualStream-RTNet provides an efficient and scalable solution for comprehensive grape quality assessment, offering strong potential for real-time sorting, precision grading, and smart agricultural applications.
Why it matches plant phenotyping methodsRGB-HSV画像とハイパースペクトルを統合し、ブドウ果実のSSCを非破壊推定する新規モデルを開発・検証しており、果実形質の取得・推定法が研究の中心である。
abstractThis study proposes DualStream-RTNet, a unified multimodal deep learning framework that simultaneously performs grape cultivar classification and SSC prediction by integrating RGB-HSV fused images and PCA-compressed hyperspectral spectra.
Accurate delineation of crop growth stages under real-world field conditions remains a long-standing challenge in computational phenotyping, particularly for wheat whose developmental phases are characterized by subtle, continuous morphological transitions and environmental noise. In this study, we propose AMFR-Net, an Adaptive Multi-Scale Feature Refinement Network tailored for fine-grained wheat stage identification using ground-level RGB imagery. Unlike conventional architectures that struggle with ambiguous inter-stage boundaries and rigid receptive structures, AMFR-Net leverages a ResNet-101 backbone augmented by a novel Adaptive Multi-Scale Attention Fusion (AMSAF) module-comprising cross-scale interaction blocks and confidence-weighted feature aggregation-to hierarchically recalibrate spatial-semantic representations. This design enables the network to adaptively amplify phenologically salient cues while suppressing irrelevant context, ensuring robust generalization under constrained annotation and deployment conditions. Evaluated on the expert-labeled CGIAR benchmark, AMFR-Net achieves state-of-the-art performance across all major metrics (Top-1 Accuracy: 89.10%; Macro-F1: 89.10%; AUC: 97.88%) and demonstrates superior discriminability in phenologically adjacent stages compared to lightweight and deep CNN baselines. Ablation studies validate the synergistic effect of multi-level attention and scale-aware refinement. The proposed framework offers a scalable, interpretable, and field-deployable solution for in-situ phenology monitoring, and sets a foundation for future integration of multimodal sensing, weak supervision, and cross-seasonal adaptation.
Why it matches plant phenotyping methods小麦の生育ステージを地上RGB画像から推定する新規深層学習手法を開発し、ベンチマーク、比較、アブレーションで検証しており、植物フェノタイピング手法が研究の中心です。
abstractwe propose AMFR-Net, an Adaptive Multi-Scale Feature Refinement Network tailored for fine-grained wheat stage identification using ground-level RGB imagery.
Reproduction assets foundThe paper's phenotyping analysis is built on the public CGIAR Wheat Growth Stage Challenge dataset (ground-level RGB wheat images with growth-stage labels), which the authors explicitly state is publicly available on Zindi with a direct link. No author analysis code, trained model checkpoints, or supplementary code/deeDataset · publicPublicly available datasets were analyzed in this study. This data can be found here: Dataset name: CGIAR Wheat Growth Stage Challenge Primary repository: Zindi (official competition page) Direct link: https://zindi.africa/competitions/cgiar-wheat-growth-stage-challengeAccession/Open asset ↗Zindilines:808-824Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
Seed selection constitutes the initial and one of the most critical steps in agricultural productivity. The identification of high-quality seeds is a labor-intensive and costly process that requires considerable expertise. Within the scope of smart farming applications, this study proposes a deep learning–based model designed to automate the seed selection process by accurately predicting seed germination capacity from seed images. The proposed model determines whether a seed will germinate using RGB images and morphological traits automatically extracted from these images. The dataset used in this study comprises a total of 3,645 images belonging to three different seed types. For each seed type (okra, eggplant, and tomato), 405 seed images were acquired from three distinct imaging sources (digital microscope, camera, and scanner), labeled, and subsequently sown in seed trays. The germination status of each sown seed was systematically monitored and matched with its corresponding image data. The dataset was partitioned into 80% training and 20% testing subsets. Following 5-fold stratified cross-validation, the proposed model achieved an average weighted F1-score of 0.95 on the training set and 0.93 on the testing set for germination capacity prediction. The performance of the proposed model was further compared with widely used deep convolutional neural network architectures, including VGG19, ResNet50, and EfficientNetB5. Comparative results demonstrate that the proposed model provides competitive and robust performance for seed germination prediction. Overall, the findings indicate that the proposed approach can effectively be utilized for automated seed germination prediction. Future research should evaluate the generalizability of the model by conducting performance assessments on additional seed types.
Why it matches plant phenotyping methodsRGB画像から種子の形態形質を自動抽出し、発芽状態を予測する深層学習手法が研究の中心であるため、植物フェノタイピング手法として含める。
abstractthis study proposes a deep learning–based model designed to automate the seed selection process by accurately predicting seed germination capacity from seed images.
Leaf morphology in tea plants (Camellia sinensis L.) profoundly influences tea quality and agronomic value, yet its genetic basis remains elusive due to labor-intensive phenotyping, foliage architecture, and ecological sensitivity of traits. Moreover, traditional methods forfeit quantitative color gradients and population-level morphological complexity. To address this challenge, we developed coleaf, an open-source image recognition-based software that demonstrated 97.6% accuracy over conventional ImageJ measurements, while offering higher efficiency and color hues quantification. We then estimated 7 key morphological traits focusing on leaves from a collection of ~ 4,200 mature leaves and ~ 5,000 bud-leaf samples across 167 genetically diverse tea accessions by coleaf. While classical understanding suggests leaf shape differentiation between two varieties in genus sinensis assamica (CSA) and sinensis (CSS), our phenotypic clustering revealed incomplete congruence with phylogenetic relationships, suggesting the presence of additional genetic or environmental modulators beyond population divergence. Furthermore, we integrated phenotypic data with whole-genome resequencing for multi-model genome-wide association studies (GWAS). Candidate genes associated with leaf architecture were involved in plant development (e.g., CsFAS2), cell division and elongation (e.g., CsFIP1), and cellular morphogenesis (e.g., CsRLK), whereas those associated with leaf color, regulated pigment accumulation (e.g., ABC transporters, CsMYB113). In conclusion, this study establishes a standardized computational framework validating automated image recognition for plant leaf phenomics. The end-to-end framework from high-throughput phenotyping to gene discovery provides critical genetic targets for tea breeding, demonstrating transformative potential in accelerating the genetic improvement of tea plants.
Why it matches plant phenotyping methods茶葉形態の画像認識ソフトウェアを開発・検証し、高スループットな形質抽出フレームワークとして適用しており、植物フェノタイピング手法が研究の中心である。
abstractwe developed coleaf, an open-source image recognition-based software that demonstrated 97.6% accuracy over conventional ImageJ measurements
Reproduction assets found保存済みの本文根拠を更新済みルールで再検証し、公開資産1件を確認しました。Code · publicAll codes and tools used in this study are described in Methods, coleaf is available on github (https://github.com/mengmeng-jiang/coleaf).Open asset ↗mengmeng-jiang/coleafhtml-lines:390-460Plant phenotyping relevance match · UnverifiedCrossref · checked 15 Sept 2026
Pest and pathogen pressure in potato cultivation is increasingly affecting the potato quality and yield. The Netherlands, as the largest seed potato producer around the world, is particularly threatened by blackleg disease and potato virus Y (PVY). Uncrewed aerial vehicle (UAV)-based imaging combined with machine- and deep-learning methods have shown clear potential for potato disease identification, offering advantages over conventional human inspections, which are labor-intensive, expertise-demanding, and often subjective. Most existing studies focused on RGB data and pixel-level classification, producing maps that have limited practical value for targeted removal of infected plants. Earlier work demonstrated the potential of plant-level disease detection approaches. For example, Jia[1] employed hyperspectral data (specifically the first three principal component analysis (PCA) bands) with a YOLOv5s model to distinguish the blackleg- and PVY-infected plants from healthy ones, yielding average mAP@.50 scores of 0.85 for blackleg detection and 0.82 for PVY detection. Gibson-Poole[2] applied object-based image analysis (OBIA) to detect blackleg disease with RGB imagery, achieving a total accuracy of 87%. The findings suggest that multi-modal data (combining hyperspectral and RGB imagery) hold strong potential for plant-level disease detection. We aim to identify the most informative features derived from hyperspectral data and to investigate their integration with RGB data to enhance potato disease detection performance.We proposed early fusion (E), where data were concatenated channel-wise before network input, and middle fusion (M) architectures, where features were extracted separately within a two-branch network and then merged at an intermediate stage, to integrate hyperspectral features and RGB imagery for potato disease detection. To reduce hyperspectral dimensionality, two feature sets were extracted: (i) the first three PCA bands, and (ii) 10 vegetation indices (VIs) selected from 64 candidates using variance inflation factor analysis to mitigate multicollinearity. Consequently, four models were developed and evaluated: E-PCA-RGB, E-VI-RGB, M-PCA-RGB, and M-VI-RGB. Unlike previous studies that focused on a single disease, our models detected blackleg-infected, PVY-infected, and healthy plants simultaneously. E-VI-RGB achieved the highest mAP@.50 value of 86.65±1.53, followed by M-VI-RGB (85.74±1.75). E-PCA-RGB and M-PCA-RGB yielded mAP@.50 scores of 83.00±2.52 and 83.11±2.20, respectively. These results demonstrate that combining hyperspectral features with RGB imagery improves detection performance compared with single-modality approaches (RGB 83.21±1.31, PCA 79.71±1.30, VIs 85.31±2.11). Our findings highlight the potential of multimodal fusion for potato disease detection in practice. The methods could enable automated systems not only to identify infected plants but also to support timely removal with machinery, mitigating the spread of disease in potato fields. The generalizability of our approach will be further tested and analyzed in future work.References[1] Jia, T., Smigaj, M., Kootstra, G. and Kooistra, L., 2024. Detection of Diseased Potato Plants with UAV Hyperspectral Imagery. In 2024 14th Workshop on Hyperspectral Imaging and Signal Processing: Evolution in Remote Sensing (WHISPERS) (pp. 1-5). IEEE.[2] Gibson-Poole, S., Humphris, S., Toth, I. and Hamilton, A., 2017. Identification of the onset of disease within a potato crop using a UAV equipped with un-modified and modified commercial off-the-shelf digital cameras. Advances in Animal Biosciences, 8(2), pp.812-816.
Why it matches plant phenotyping methodsUAVのハイパースペクトル・RGB画像からジャガイモ個体の病害状態を推定するマルチモーダル融合手法を開発・評価しており、植物表現型の取得・抽出が研究の中心です。
abstractWe proposed early fusion (E), where data were concatenated channel-wise before network input, and middle fusion (M) architectures, where features were extracted separately within a two-branch network and then merged at an intermediate stage, to integrate hyperspectral features and RGB imagery for potato disease detection.
Reliable agricultural statistics support food security monitoring and evidence-based decision making. In Mozambique, official agricultural statistics are primarily derived from the Integrated Agricultural Survey (IAI), an enumerator-based field survey that provides essential contextual information on agricultural production but remains labour-intensive, costly and spatially and temporally constrained, particularly in remote rural areas. While satellite remote sensing offers complementary, wall-to-wall coverage, its spatial resolution is often insufficient to directly capture the fragmented fields, mixed and intercropping patterns, shifting cultivation and strong sub-field variability typical of smallholder farming systems. Consequently, consistent estimation of crop area and crop type derived from enumerator-based crop cover assessments remains challenging in these landscapes.This study investigates the potential of high-resolution multispectral data acquired with Uncrewed Aerial Vehicles (UAVs) to complement field surveys by providing spatially explicit and internally consistent crop cover and crop fraction estimates at the field and sub-field scale. By resolving individual crops and dominant intercropping systems, UAV-based observations support the interpretation of farmer-reported crop cover proportions, improve consistency across enumerators, and enable post-survey correction of crop area estimates, while providing a basis for future integration with coarser-resolution satellite remote sensing. High-resolution RGB and multispectral imagery (green, red, red edge, and near-infrared; ≤5 cm ground sampling distance) was collected using a DJI Mavic 3M with RTK over 30 sampling areas of 500 × 500 m in Manica Province during the 2025 agricultural season. In parallel, a field survey recorded standardized observations of agricultural activity, including crop type (of most field and tree crops), intercropping combinations and enumerator-based estimates of fractional crop cover. UAV images were processed using a workflow tailored to heterogeneous smallholder landscapes to produce orthomosaics, digital surface models (DSMs), and vegetation indices. These products were linked to field observations through segments representing relatively homogeneous land units, enabling direct comparison between UAV-derived and survey-based crop cover estimates.For crop classification, training polygons were delineated on RGB orthomosaics for single-crop fields (e.g. maize, beans, sorghum and cassava) and common intercropping combinations (e.g. maize–beans). Annotated mosaics were tiled and augmented and used to train convolutional neural network models (e.g. UNet++), incorporating multispectral vegetation indices and DSM-derived height information as additional input channels. Model performance was evaluated using Intersection over Union, Dice coefficients, and regression metrics for fractional cover accuracy.A comparison framework was implemented to relate UAV-derived crop type, crop combinations and fractional cover to field survey observations while explicitly accounting for measurement uncertainty. Model II regression quantified systematic bias and proportional differences between the two methods. Initial results indicate that UAV-derived estimates provide spatially consistent crop cover information in fields with complex intercropping structures. Ongoing work focuses on refining segmentation accuracy, analysing residual discrepancies and assessing how UAV-derived crop cover information can be integrated to expand the spatial coverage and reliability of agricultural statistics in smallholder landscapes.
Why it matches plant phenotyping methodsUAVマルチスペクトル画像から作物種・混植構造・作物被覆率を抽出し、CNN分類と分画被覆推定を検証する方法が研究の中心である。
abstractUAV-based observations support the interpretation of farmer-reported crop cover proportions, improve consistency across enumerators, and enable post-survey correction of crop area estimates
Abstract The major aim of the study is to determine some of the diseases affecting rice plants which lead to crop loss. The suggested model is designed into four primary phases, i.e., pre-processing, feature extraction, feature selection, and classification. The proposed models have been applied to two sets of data that include Rice Disease and the Rice Leaf Disease Image that constitute a total of four types of classes of the paddy leaves: healthy, blast, bacterial blight, and tungro. The dataset images are initially upgraded in the first stage of pre-processing in order to enhance the quality of the images. A Gaussian filter is used to eliminate noise on the green spectral band and to convert the input images to RGB color space. The step second involves deriving color and texture information out of each of the pre-processed images. The third step Features selection SVM-RFE has been used to select features with the help of the intersection with the ARFA technique. The fourth step involves a final step where the features chosen are applied to classify the kind of disease existing in each picture. ABi-LSTM model is used in the classification process and the accuracy is 97.05.
Why it matches plant phenotyping methodsイネ葉画像から病害状態を推定する画像ベースの表現型解析手法を、前処理・特徴抽出・特徴選択・分類のワークフローとして開発しており、方法が中心である。
abstractThe suggested model is designed into four primary phases, i.e., pre-processing, feature extraction, feature selection, and classification.
BACKGROUND: Root biomass serves as a critical indicator of plant eco-physiological status and crop productivity, yet its non-destructive monitoring remains challenging because of its underground location. The use of transparent nutrient film technique (NFT) systems enables direct observation of entire root systems, rendering image-based phenotyping feasible. In this study, we investigated and compared the performance of RGB and hyperspectral imaging for predicting root dry weight in hydroponically grown spinach (Spinacia oleracea L.). RESULTS: Using 430 root segments divided from 60 plants, three models were developed: (1) an area-based regression based on root coverage, (2) a convolutional neural network (CNN) using RGB images, and (3) a partial least squares regression (PLSR) model using hyperspectral data (450-950 nm). The area-based regression exhibited limited accuracy (R² = 0.446) because of saturation at high root coverage. The CNN model improved predictive performance (R² = 0.739) but tended to overestimate sparse roots as a result of resolution constraints. The PLSR model achieved the highest accuracy (R² = 0.822, RMSE = 0.019 g/segment), with significantly lower error than RGB-based approaches (P < 0.01). Variable importance in projection analysis indicated that PLSR effectively exploited spectral signatures at 450 nm (background contrast) and 750 nm (tissue scattering), thereby maintaining stable accuracy across the full biomass range. When validated using 104 independent plants, the PLSR model achieved high predictive accuracy. Furthermore, as a proof of concept, this model successfully visualized the spatiotemporal dynamics of root biomass accumulation over 50 days, with only a 7.70% relative error at harvest. CONCLUSIONS: To our knowledge, this study is among the first to demonstrate the non-destructive monitoring of biomass distribution within entire root systems under production conditions. Hyperspectral imaging combined with PLSR outperforms RGB-based approaches by capturing spectral signatures that reflect internal tissue properties of roots, thereby overcoming limitations caused by morphological occlusion. This approach provides a robust tool for precision agriculture and high-throughput phenotyping, enabling continuous assessment of root growth through simple modifications to the existing hydroponic systems.
Why it matches plant phenotyping methodsRGB・ハイパースペクトル画像と機械学習／PLSRを用いて根乾物重を非破壊推定・検証する方法研究であり、植物表現型の取得と定量化が中心である。
abstractThe use of transparent nutrient film technique (NFT) systems enables direct observation of entire root systems, rendering image-based phenotyping feasible.
Reproduction assets foundThe paper's Data availability statement deposits the paper-specific phenotyping assets (raw hyperspectral images, RGB images, and root dry weight measurements) in a Zenodo record. The provided URL includes a token and 'preview=1', suggesting the record may not yet be fully open, but it is the authors' stated public URLDataset · publicThe datasets generated and analyzed during the model construction of the current study are available in the Zenodo repository: [https://zenodo.org/records/18072801?preview=1&token=eyJhbGciOiJIUzUxMiJ9.eyJpZCI6ImFiYTkzMzY2LTIzZjktNDlkMy1iZTBjLTk3M2E5YTUyOTFmZCIsImRhdGEiOnt9LCJyYW5kb20iOiIzNGE1ZjMxNDZhYjhiYjlhZWRiOWFjNzBkNzcwY2I3NyJ9.uR4HfosoSaVWhtSblMOS1v9bJFA5MvHwXvcW9uoNbcTWRDU4RNxZpVHjXTC3ulBM1JTlBbeHp_4T5EcILawxdg].The dataset includes:
- Raw hyperspectral images and data- RGB images
- Root dry weight measurementsOpen asset ↗Zenodo · 18072801lines:176-248Code / dataset availability confirmedEurope PMC · checked 15 Sept 2026
Service crops are grown to provide ecosystem services in viticulture, but their adoption remains limited due to their competition with grapevine for soil resources. To identify trade-offs between services, the effect of service crops management strategies on grapevine performances still need further research. This dataset presents data from two experiments conducted to study the effect of service crops management on soil resources and grapevine performances. The inter-row vegetation was sampled in two Mediterranean vineyards using quadrats for biomass estimation. In addition, an unmanned aerial vehicle (UAV) was regularly flown over the vineyards for a period spanning more than four years in total over the two vineyards. The dataset presented here includes both raw data acquired during fieldwork and processed data derived from this raw inputs. The raw data consists of image series captured by two UAVs during each flight campaign, including RGB and multispectral imagery. Images were acquired between 2021-06-10 and 2022-07-29 for the first vineyard, and between 2023-06-08 and 2025-03-12 for the second vineyard. Based on these raw data, the processed data comprises spatial vectors, raster layers, and dense point clouds generated from UAV images using a Structure from Motion (SfM) photogrammetry workflow, at a 5 cm spatial resolution. The raster layers and dense point clouds provide specific information on vineyard characteristics for each UAV flight date, including elevation, vegetation indices, visible and near-infrared reflectance, and canopy height. In addition, the processed data include measurements of vegetation dry biomass, as well as separate measurements of dry biomass and leaf area measured for selected service crops species. This dataset can be reused for the calibration and/or evaluation of classification algorithms aimed at discriminating vines from the inter-row vegetation, or as part of a larger dataset to explore relationships between remotely-sensed vegetation indices and field-measured vegetation biomass or surface.
Why it matches plant phenotyping methodsUAV画像とSfM処理により、植生指数・樹冠高・バイオマス等の植物形質を取得した再利用可能なデータセットで、分類アルゴリズムの校正・評価用途も明示されており、植物フェノタイピング手法・データ基盤が中心です。
abstractThe dataset presented here includes both raw data acquired during fieldwork and processed data derived from this raw inputs.
Reproduction assets foundThe paper is a Data in Brief article describing a public dataset on Research Data Gouv (doi: 10.57745/MXM55R) containing UAV RGB/multispectral imagery, SfM-derived rasters and point clouds, and field-measured vegetation biomass/leaf-area data from two Mediterranean vineyards — directly the paper's phenotyping inputs. ADataset · publicollected in vineyards located in southern France near Montpellier (43°32.5243′N, 3°50.8240′E). Data are stored on Research Data Gouv, a remote storage solution curated by the French Department of Research.
Data accessibility
Repository name: Research Data Gouv
Data identification number: doi: 10.57745/MXM55R
Direct URL to data: https://doi.org/10.57745/MXM55R
Related research article
None
1.
Value of the Data
•
The fine scale imaging of vineyards (i.e., 5 cm resolution) allows for classification of the vegetation in the vineyard inter-rows, and subsequent exploration of its respective dynamics.
•Open asset ↗Research Data Gouv · 10.57745/MXM55Rlines:1-47Plant phenotyping relevance match · UnverifiedEurope PMC · checked 5 Sept 2026
Breeding rice varieties that are both salt-tolerant and high-yielding is essential for utilizing saline-alkaline lands and ensuring food security. However, However, high-throughput and accurate phenotyping at early growth stages remains a major bottleneck in breeding programs. In this study, unmanned aerial vehicle (UAV) imaging was employed to screen salt-tolerant and high-yielding varieties among 60 rice varieties under saline-alkaline field conditions. Red-green-blue (RGB), multispectral, and thermal canopy images were acquired throughout the growing season by UAV, from which 41 phenotypic traits were extracted at each growth stage. These traits were categorized into early-stage (tillering and jointing), late-stage (booting, flowering, and maturity), and whole-growth-stage (from tillering to maturity) and subsequently used to screen salt-tolerant and high-yielding rice varieties. Results showed that: (1) An early high-throughput screening method for salt-tolerant rice varieties was developed based on the membership function and UAV phenotypes (MFuav), achieving high performance (Precision >0.8, OA > 0.7). MFuav demonstrated the highest accuracy at the early-stage, with Precision increasing by 0.29 and 0.43 compared to the late- and whole-stage models, respectively. (2) A machine learning based UAV phenotypes framework (MLuav) was developed to further improve salt-tolerance screening performance. Within this framework, the partial least squares regression (PLSR) was employed for early-stage salt-tolerance screening, which achieved a Precision of 0.97 and an OA of 0.78, outperforming the MFuav by 0.11 and 0.08, respectively. In addition, within the same MLuav framework, early-stage UAV phenotypes were further used for actual yield prediction using a Random Forest (RF) model. The model achieved a high Recall for high-yielding varieties (Recall = 1.00), ensuring that no potentially high-yielding germplasm was missed, although this was accompanied by a moderate Precision (0.51) and an overall accuracy of 0.70. (3) The MLuav consistently outperformed the MFuav in screening salt-tolerant and high-yielding varieties across all 60 rice varieties. Among the five referenced salt-tolerant and high-yielding rice varieties, the MLuav correctly screened four using early-stage phenotypes, whereas the MFuav only screened three. Overall, the proposed method enables early screening of salt-tolerant and high-yielding rice varieties, offering an efficient tool for the screening and utilization of elite stress-resilient germplasm.
Why it matches plant phenotyping methodsUAV画像から多数のイネ表現型形質を抽出し、塩耐性・収量性を早期スクリーニングする方法と機械学習フレームワークを開発・評価しており、表現型取得・解析手法が研究の中心である。
abstracthigh-throughput and accurate phenotyping at early growth stages remains a major bottleneck in breeding programs
Rice early tillering characteristics are key indicators for high-yield breeding, with tiller number and tillering rate as core parameters. High-throughput, temporal, and precise monitoring of tiller numbers via drone digital imagery provides quantitative support for tillering trait screening in breeding, serving as an important auxiliary tool for smart breeding. However, during the early tillering stage, complex backgrounds (e.g., water bodies, soil) and small, dense breeding plots pose challenges to high-throughput rice plant extraction and accurate tiller number estimation. To address this, this study proposes a rice tiller number estimation method based on an improved Swin-UNet model and multi-feature fusion. A PSO-optimized XGBoost model was constructed for tiller number estimation by integrating selected features. Experimental results show that the improved Swin-UNet model achieved a segmentation accuracy of 92.5% (7.2% higher than U-Net), and the PSO-XGBoost model, using 12 features (10 morphological and 2 color), yielded R²=0.85 and RMSE = 0.35. Application verification on 576 untrained breeding plots generated tiller number thematic maps, providing data support for germplasm tillering trait identification and advancing smart breeding.
Why it matches plant phenotyping methodsドローン画像からイネの分げつ数を抽出・推定する画像解析手法を開発し、セグメンテーション精度と推定性能を検証しているため、植物表現型計測が研究の中心である。
abstractthis study proposes a rice tiller number estimation method based on an improved Swin-UNet model and multi-feature fusion.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 5 Sept 2026
This study addresses the challenge of automated high-throughput phenotyping of wheat spike characteristics using modern computer vision and deep learning methods. Accurate estimation of spikelet number is a key indicator of plant productivity, yet traditional manual counting approaches are labor-intensive, slow, and difficult to scale to large breeding datasets. To overcome these limitations, we propose a spikelet detection strategy based on simplified point annotations, where an expert marks only the centers of spikelets rather than drawing detailed segmentation masks or bounding boxes. This significantly reduces annotation time and lowers the overall cost of preparing training datasets for machine learning models. To determine the most effective way of utilizing such simplified annotations, three computational methods were explored: segmentation of binary masks using a U-Net architecture, density regression based on two-dimensional Gaussian distributions optimized via Kullback-Leibler divergence, and detection of fixed-size bounding regions using the YOLOv8 object detection framework. The models were evaluated on dedicated test datasets using both quantitative metrics (MAE, MAPE) and spatial localization metrics (Precision, Recall, F1 score). The results demonstrate that U-Net-based approaches provide consistently high accuracy in spikelet localization and counting while maintaining robustness to annotation imperfections. In contrast, the YOLOv8-based method showed reduced performance, likely due to the geometric mismatch between fixed-size boxes and the natural elongated shape of spikelets. Overall, the proposed methodology highlights the effectiveness of combining minimalistic point-level annotation with advanced segmentation models for automating phenotyping workflows. This approach has the potential to accelerate breeding programs, enhance the efficiency of large-scale phenotypic data collection, and support further development of robust computer-vision tools for plant science applications.
Why it matches plant phenotyping methodsコムギの穂の小穂数をRGB画像から自動推定する画像解析・深層学習手法を開発し、複数モデルを定量評価しており、フェノタイピング手法が研究の中心である。
abstractThis study addresses the challenge of automated high-throughput phenotyping of wheat spike characteristics using modern computer vision and deep learning methods.
Abstract Cherry is widely cultivated but remains challenging to harvest due to economic and ecological constraints, especially in developing countries such as Pakistan. Climate change, limited use of technology, and foliar diseases worsened by pesticide use further reduce productivity, particularly during fruiting. Conventional disease assessment depends on expert observation and grower experience, making it subjective and time-consuming. A comprehensive evaluation was conducted on the PlantCity dataset, which contains 5,714 high-density, full-color RGB images collected under challenging conditions and categorized into 5 classes. We compared three approaches: deep learning pre-trained, transfer learning, and a machine learning pipeline. Models were evaluated by accuracy, precision, recall, F1-score, Cohen’s Kappa, inference time, FLOPs, and throughput. Grad-CAM was used to improve interpretability. Transfer learning using DenseNet169 achieved the highest performance, with 99.80% accuracy, 99.80% precision, 99.80% recall, and a Cohen’s Kappa of 99.74%. These results were significantly higher than those obtained by other deep learning architectures and handcrafted baselines. Grad-CAM heatmaps confirmed that the models focused their attention on pathological areas. The proposed transfer-learning-based framework, particularly DenseNet169, demonstrates state-of-the-art diagnostic accuracy and features a modular structure. This design enables deployment on both high-performance servers and resource-constrained embedded devices, thereby facilitating early disease detection in precision agriculture.
Why it matches plant phenotyping methods植物画像から病害状態を分類する画像ベース表現型推定が研究の中心であり、複数モデルの比較・性能評価とGrad-CAMによる検証を行っているため。
abstractConventional disease assessment depends on expert observation and grower experience, making it subjective and time-consuming.
Intelligent forest tree breeding has advanced plant phenotyping, yet existing research largely focuses on large-leaf agricultural crops, with limited attention to fine-grained leaf analysis of sapling trees in open-field environments. Natural scenes introduce challenges including scale variation, illumination changes, and irregular leaf morphology. To address these issues, we collected UAV RGB imagery of field-grown saplings and constructed the Poplar-leaf dataset, containing 1,202 branches and 19,876 pixel-level annotated leaf instances. To our knowledge, this is the first instance segmentation dataset specifically designed for forestry leaves in open-field conditions. We propose LeafInst, a novel segmentation framework tailored for irregular and multi-scale leaf structures. The model integrates an Asymptotic Feature Pyramid Network (AFPN) for multi-scale perception, a Dynamic Asymmetric Spatial Perception (DASP) module for irregular shape modeling, and a dual-residual Dynamic Anomalous Regression Head (DARH) with Top-down Concatenation decoder Feature Fusion (TCFU) to improve detection and segmentation performance. On Poplar-leaf, LeafInst achieves 68.4 mAP, outperforming YOLOv11 by 7.1 percent and MaskDINO by 6.5 percent. On the public PhenoBench benchmark, it reaches 52.7 box mAP, exceeding MaskDINO by 3.4 percent. Additional experiments demonstrate strong generalization and practical utility for large-scale leaf phenotyping.
Why it matches plant phenotyping methods森林葉の個体分割と表現型解析のためのUAV画像データセットおよび新規セグメンテーション手法を開発・評価しており、植物表現型取得が中心である。
abstractwe collected UAV RGB imagery of field-grown saplings and constructed the Poplar-leaf dataset, containing 1,202 branches and 19,876 pixel-level annotated leaf instances.
Plant pathogens cause yield losses worldwide, threatening food security and livelihoods. Because early infection is difficult to diagnose, management often relies on prophylactic pesticide use, increasing costs and environmental impact. Here we present PSNet, a multimodal framework that fuses hyperspectral imaging with RGB information for presymptomatic plant disease detection, together with a low-cost hyperspectral camera incorporating a 3D-printed housing, costing under £500. We validate PSNet using Arabidopsis thaliana infected with the oomycete Albugo candida . Imaging at 2 and 4 days post inoculation, prior to visible symptoms, revealed spectral signatures that distinguished infected from healthy plants, while imaging at 6 days post inoculation captured the transition toward early symptom emergence. Discriminative spectral regions overlapped wavelengths associated with plant responses to biotic stress, supporting the biological plausibility of these signatures. Performance was evaluated using strict plant-level partitioning, ensuring samples from the same plant were confined to a single split. On a four-class task (healthy, 2 dpi, 4 dpi, 6 dpi), PSNet achieved 90.00% accuracy and 97.50% accuracy for binary classification. Together, these results demonstrate that presymptomatic detection is feasible under controlled conditions using low-cost hardware and multimodal learning, underscoring the potential of scalable multimodal systems for early disease monitoring.
Why it matches plant phenotyping methods低コストのハイパースペクトル・RGB融合による植物病害状態の非破壊推定手法を開発し、植物単位で性能検証しているため、フェノタイピング手法が中心である。
abstractHere we present PSNet, a multimodal framework that fuses hyperspectral imaging with RGB information for presymptomatic plant disease detection, together with a low-cost hyperspectral camera incorporating a 3D-printed housing, costing under £500.
Early Disease Detection (EDD) in plants is crucial for identifying infections before serious symptoms manifest, ensuring agricultural product safety, reducing chemical use, improving orchard hygiene, and providing a sustainable disease control method. The research developed a EDALMG model to improve apple disease recognition and orchard productivity by combining RGB and multispectral imaging. It segments data from Kaggle datasets for training and creates vegetation indices to highlight plant physiology. The model uses a data-level fusion approach, channel attention, and hierarchical feature extraction through ResNet for comprehensive disease features. It employs multi-label classification to identify multiple diseases simultaneously, enhancing system robustness. The Gaussian bounding method is preferred for tracking infection areas due to its precise 2D positional data usage. The performance analysis includes Loss Calculation, Accuracy Calculation, Confusion Matrix Calculation, and F1-Score Calculation.
Why it matches plant phenotyping methodsリンゴ葉・果実の感染領域と病害状態をRGB・マルチスペクトル画像から推定する分類・位置推定手法が研究の中心であり、植物病害フェノタイピングに該当する。
abstractThe research developed a EDALMG model to improve apple disease recognition
Accurate flower-load assessment is critical for informed thinning strategies in orchard management. UAV-based deep learning automated counting offers efficiency advantages, yet precise counting is heavily dependent on abundant annotated data, which is scarce and costly to obtain in agricultural settings. While semi-supervised learning alleviates dependency on manual annotation, its application to UAV-based orchard imagery faces challenges: complex backgrounds and small target sizes, which undermine pseudo-label reliability. To address these challenges, this study proposes a two-stage framework to achieve separate counting of apple flowers at different phenological stages. First, a color-SAM flower extractor (CSAM-FE) is proposed to preprocess images using a strategy combining color thresholding with the Segment Anything Model (SAM), suppressing background noise and extracting high-quality flower clusters, thereby providing purified inputs for the subsequent counting network. Second, an uncertainty-guided semi-supervised flower counting network (USCount-Net) is proposed for accurate stage-specific flower counting with limited labeled data. The USCount-Net incorporates two key components: an adaptive pseudo-label filtering (PLF) mechanism based on frequent forward uncertainty estimation (FFUE) is designed to dynamically suppress noisy gradient backpropagation, mitigating error propagation from unreliable pseudo-labels; and a noise-sensitive adaptive gated fusion (AGF) module is introduced to fuse cross-scale features without redundancy, addressing significant scale variations across phenological stages and observation angles. Comparative experiments on a self-built apple flower counting dataset demonstrate that USCount-Net achieves lower MAE and RMSE than state-of-the-art methods at 10%, 30%, and 50% labeling ratios. The results demonstrate that the proposed methodology serves as methodological support for rapid and precise apple flower counting in low-annotation agricultural scenarios.
Why it matches plant phenotyping methodsリンゴ花の画像抽出・計数手法と半教師あり解析ネットワークを開発し、データセット上で比較評価しているため、植物表現型取得が中心である。
abstractthis study proposes a two-stage framework to achieve separate counting of apple flowers at different phenological stages.
Reproduction assets foundThe paper's Data availability statement explicitly provides public access to the authors' USCount-Net source code on GitHub and the self-built apple flower counting dataset (UAV images, annotations, flower cluster images) on Google Drive.Code · publicThe source code is publicly available at https://github.com/haohuihui5019/USCount-Net . And the source dataset can be accessed at https://drive.google.com/drive/folders/1KP8H0qIuct56hWre5GV6ZJnzwOpen asset ↗USCount-Netlines:681-780Dataset · publicThe source code is publicly available at https://github.com/haohuihui5019/USCount-Net . And the source dataset can be accessed at https://drive.google.com/drive/folders/1KP8H0qIuct56hWre5GV6ZJnzwOpen asset ↗lines:681-780Code / dataset availability confirmedEurope PMC · checked 5 Sept 2026
Volume is an important shape descriptor in postharvest quality evaluation and breeding programs of sweetpotatoes and is also valuable for other agricultural engineering applications. Traditional volume measurement methods based on water displacement are, however, laborious, destructive, and unsuitable for high-throughput online scenarios. To address this gap, this dataset was developed to support the advancement of non-destructive, automated online volume estimation using a LiDAR (light detection and ranging)-based three-dimensional (3-D) machine vision system. A total of 200 sweetpotato storage roots of the cultivar "Beauregard" were collected for constructing a 3-D multi-view imagery dataset. Each sample was imaged online using a short-range LiDAR camera (Intel RealSense™ L515) while traveling on a custom-built roller conveyor system that enables simultaneous translation and rotation for full-surface coverage. The curated dataset comprises raw color images (1280 × 720 pixels, .png format) and corresponding raw and segmented point clouds (1280 × 720 pixels, .laz format) for individual samples, alongside the reference volume measurements obtained using the standard water displacement method. In addition, to illustrate the modeling pipeline for volume prediction, the dataset provides the extracted geometric features derived from the segmented two-dimensional (2-D) masks and point clouds, and volume prediction results obtained through regression modeling. As the first publicly available LiDAR-based dataset for sweetpotato volume estimation, this dataset provides a valuable resource for developing and validating image processing pipelines, optimizing machine learning models, and advancing 3-D vision technologies for non-destructive, rapid measurement of the volume of irregularly shaped agricultural products.
Why it matches plant phenotyping methodsサツマイモ貯蔵根の体積という植物器官形質をLiDAR 3D画像から推定する公開データセットであり、取得系・参照測定・特徴抽出・予測結果を含むため、フェノタイピング手法とデータセットが中心です。
abstractthis dataset was developed to support the advancement of non-destructive, automated online volume estimation using a LiDAR (light detection and ranging)-based three-dimensional (3-D) machine vision system.
Reproduction assets foundThe paper's own LiDAR sweetpotato dataset (images, point clouds, ground-truth volumes, feature data, and Python modeling scripts) is publicly deposited on Zenodo with an explicit DOI. The librealsense GitHub link is a generic camera SDK, not a paper-specific asset.Dataset · publicDirect URL to data: https://doi.org/10.5281/zenodo.18378019Open asset ↗Zenodo · 10.5281/zenodo.18378019html-lines:90-113Code · publicThe complete Python modeling script and the associated feature datasets have been included in the public dataset repository [13] to facilitate reproducibility and provide a benchmark for future algorithm development.Open asset ↗html-lines:168-182Code / dataset availability confirmedEurope PMC · checked 5 Sept 2026
The wheat canopy genome harbors abundant yet untapped genetic variation that could be harnessed to enhance yield potential. The green area index (GAI) is a structural metric that reflects the photosynthetically active canopy surface and is closely linked to final grain yield. Current image-based GAI retrieval methods often suffer from signal saturation and coarse structural depiction, constraining downstream genetic analyses. To address this limitation, we constructed a comprehensive image dataset spanning eight field experiments across China and France, encompassing approximately 600 genotypes under six distinct management regimes. Leveraging this diverse data, we developed a multimodal deep-learning framework augmented by simulated-to-realistic (sim2real) synthetic data transfer. This framework fuses nadir and oblique RGB images with accumulated thermal time to produce high-precision, time-series GAI estimates. Validated on independent testing datasets from both China and France, the multimodal approach demonstrated robust performance with an accuracy of R 2 = 0.88 and an RMSE of 0.49 m 2 m -2 , representing an improvement of about 22% over the traditional gap fraction method. In three site-year field experiments involving 565 genotypes, the GAI dynamics derived from the multimodal approach showed higher broad-sense heritability (0.20-0.48) than those from the gap fraction approach (0.02-0.13) and stronger genotypic correlations with yield (0.19-0.40 versus 0.09-0.31). Furthermore, genetic analysis confirmed the biological fidelity of the estimated traits, identifying loci that co-localize with known architectural regulators such as Rht-D1 , TaTB1-4D , and TaBGC1-4D . Consistently, the multimodal-derived phenotypes were specifically enriched in cell-wall remodeling and hormonal signaling pathways (e.g., brassinosteroid) that directly regulate canopy expansion. Overall, the proposed method offers a powerful tool for unlocking genetic gain in canopy architecture and accelerating canopy-targeted wheat improvement.
Why it matches plant phenotyping methodsデュアルアングルRGB画像と熱時間を統合してGAIを推定する深層学習法を開発し、独立データで検証しているため、植物形質取得法が研究の中心です。
abstractwe constructed a comprehensive image dataset spanning eight field experiments across China and France
Reproduction assets foundThe paper publicly releases its pre-trained multimodal GAI-estimation model weights and inference code on Hugging Face, directly reproducing this paper's phenotyping analysis. The raw image and phenology datasets are not public and require contacting the authors.Code · publicThe pre-trained model weights, inference code, and usage instructions are publicly available in the Hugging Face repository at https://huggingface.co/PheniX-Lab/GAI-Estimation/tree/main .Open asset ↗PheniX-Lab/GAI-Estimationlines:259-277Plant phenotyping relevance match · UnverifiedCrossref · OpenAlex · checked 15 Sept 2026
NeRF / 3D Gaussian SplattingLiDAR / point cloudRGB / grayscaleWhole plant / canopy / plot / fieldOrgan identification2D/3D reconstructionSegmentation
Organ segmentation of plant point clouds is a prerequisite for the high-resolution and accurate extraction of organ-level phenotypic traits. Although the fast development of deep learning has boosted much research on segmentation of plant point clouds, the existing techniques for organ segmentation still face limitations in resolution, segmentation accuracy, and generalizability across various plant species. In this study, we proposed a novel approach called plant segmentation neural radiance fields (PlantSegNeRF), aiming to directly generate high-precision instance point clouds from multi-view RGB image sequences for a wide range of plant species. PlantSegNeRF performed two-dimensional (2D) instance segmentation on the multi-view images to generate instance masks for each organ with a corresponding instance identification (ID). The multi-view instance IDs corresponding to the same plant organ were then matched and refined using a specially designed instance matching (IM) module. The instance NeRF was developed to render an implicit scene containing color, density, semantic and instance information, which was ultimately converted into high-precision plant instance point clouds based on volume density. The results proved that in semantic segmentation of point clouds, PlantSegNeRF outperformed the commonly used methods, demonstrating an average improvement of 16.1 %, 18.3 %, 17.8 %, and 24.2 % in precision, recall, F1-score, and intersection over union (IoU) compared to the second-best results on structurally complex datasets. More importantly, PlantSegNeRF exhibited significant advantages in instance segmentation. Across all plant datasets, it achieved average improvements of 11.7 %, 38.2 %, 32.2 % and 25.3 % in mean precision (mPrec), mean recall (mRec), mean coverage (mCov), and mean weighted coverage (mWCov), respectively. Furthermore, PlantSegNeRF demonstrates superior few-shot, cross-species performance, requiring only multi-view images of few plants to train models applicable to specific or similar varieties. This study extends organ-level plant phenotyping and provides a high-throughput way to supply high-quality 3D data for developing large-scale artificial intelligence (AI) models in plant science. • A comprehensive dataset of well-labeled two-dimensional (2D) images and point clouds dataset of plants was established, including various varieties and growth stages. 50 plant samples were collected for each type. • A novel multi-view image instance matching (IM) module was proposed to align plant organ instance identifications (IDs) across different viewpoints, serving as the foundation for organ-level instance segmentation. • A multi-channel instance neural radiance fields (NeRF) module with encoding color, semantic, and instance information was developed to achieve high-precision mapping of 2D image colors, semantics, and aligned instances into 3D space, enabling point cloud background removal and fine-grained segmentation of plant organs.
Why it matches plant phenotyping methods植物器官の3D点群再構成・インスタンス分割を開発し、セグメンテーション性能を検証する手法研究であり、器官レベル表現型抽出を直接支援するため。
abstractwe proposed a novel approach called plant segmentation neural radiance fields (PlantSegNeRF), aiming to directly generate high-precision instance point clouds from multi-view RGB image sequences for a wide range of plant species.
The diameter of natural rubber trees serves as a critical crop parameter, not only for determining whether rubber trees meet tapping requirements and assessing their growth status, but also playing a significant role in calculating parameters such as tapping angle, trajectory, depth and yield prediction. To further advance the intelligent production level of natural rubber trees and achieve low-cost, automated diameter measurement methods, this study proposes a new approach that combines YOLO11s instance segmentation algorithms with a monocular RGB camera to enable non-contact and non-fixed distance diameter measurement of natural rubber trees. The YOLO11-seg is used to obtain masks and bounding boxes for the ID, trunk, and tapped area. This method employs image processing techniques such as contour smoothing, trunk skeleton extraction, angle calculation, and localization. With using the tree ID tag as the primary dimensional reference and incorporating the segmentation contours of trunk categories, it achieves the measurement and calculation of rubber tree trunk diameter. The results demonstrated that among the compared instance segmentation models, the highest segmentation accuracy mAP50-95ˢᵉᵍ reached 0.934. The diameter estimation based on this segmentation and geometric correction process achieved a root mean square error (RMSE) of 2.85 cm and a mean absolute percentage error (MAPE) of 12.58 % under the original measurement conditions. After error compensation, the RMSE and MAPE decreased to 2.13 cm and 8.68 %, respectively. The proposed method can accurately measure the diameter of natural rubber trees, significantly reducing the hardware cost. It provides a new approach for measuring the diameter of natural rubber trees, and also provides both theoretical support and practical basis for the intelligent production and precision agriculture in natural rubber cultivation.
Why it matches plant phenotyping methodsゴム樹の幹径という植物形態形質を、RGB画像とインスタンスセグメンテーションで非接触・自動推定する手法を開発し、精度検証まで行っており、フェノタイピング手法が中心である。
abstractthis study proposes a new approach that combines YOLO11s instance segmentation algorithms with a monocular RGB camera to enable non-contact and non-fixed distance diameter measurement of natural rubber trees.
MaizeRiceSoybeanField / plotMultimodalLiDAR / point cloudRGB / grayscaleMultispectral / hyperspectralThermalWhole plant / canopy / plot / field
Plant phenotyping is essential for elucidating genotype–environment interactions, yet conventional methods remain labor-intensive and low-throughput. TraitDiscover transcends these constraints by uniting multimodal sensing with tightly coupled hardware-software orchestration in a single, end-to-end phenotyping platform. Aligned with the ”Plant Phenotyping Trinity” framework, the system comprises a millimetre-accurate triaxial automation unit, a modular sensor array–RGB imaging, three-dimension laser scanner or LiDAR (3D), infrad (IR) thermal imaging, hyperspectral imaging (HSI), and photosynthesis (PS) imaging–and the dedicated software TraitNavigator suite into one cohesive system. A unified spatiotemporal synchronization mechanism enables robust time-series analysis and fusion of multisource phenotypic data across the entire crop growth period, while the DepthCropSeg algorithm and a night-time imaging module enhance trait extraction under complex conditions, providing G × E × P-ready, multimodal phenotypic datasets. Validation across soybean, maize, and rice trials demonstrated high sensitivity—detecting drought stress four days before visible symptoms, identifying glyphosate injury 24 hours ahead of manual scoring, and quantifying local adaption patterns across ecological gradients. While challenges remain in scaling to complex open-field conditions, TraitDiscover offers a scalable, data-driven approach to accelerate stress phenotyping and breeding decisions and is readily poised for deeper integration with AI to advance sustainable agriculture.
Why it matches plant phenotyping methodsマルチモーダルセンシング、画像解析、同期機構、形質抽出アルゴリズムを統合した植物フェノタイピング基盤の開発と検証が中心であり、ストレス検出や形質定量も実証している。
abstractTraitDiscover transcends these constraints by uniting multimodal sensing with tightly coupled hardware-software orchestration in a single, end-to-end phenotyping platform.
Developing climate-resilient wheat varieties requires combining high yield with stability across diverse environments, especially under increasingly variable precipitation and rising temperatures. This study evaluated 64 post-Green Revolution durum wheat cultivars under irrigated and rainfed conditions at two contrasting Mediterranean sites in Spain. A classification framework was developed to support genotype selection based on yield and yield stability, estimated using linear mixed models and yield slopes across environments. Genotypes were classified by interquartile thresholds, and those showing either low yield or low stability were considered undesirable for selection. High-throughput phenotyping was conducted throughout the season using ground-sensor Red-Green-Blue (RGB) and multispectral (MS) vegetation indices (VIs), along with UAV-derived RGB, MS, and thermal-infrared (TIR) data. VIs and TIR at anthesis and grain filling, and their differences (senescence proxies), were used to train Random Forests for yield and stability estimation including sequential feature selection. Environmental covariates (water input, reference evapotranspiration) were integrated in yield models, with strong outcomes (R 2 > 0.74; MAPE <23.6%). Stability predictions were based on VI stability and, though moderate (R 2 up to 0.56; MAPE <17.75%), outperformed previous studies. Selected features were used to evaluate seasonal reflectance phenotypes: “keep” genotypes (intermediate/high yield or/and stability) exhibited early-vigor but lower green retention by the end of grain filling, while “discard” genotypes (low yield or/and stability) showed reduced early vigor and “stay-green” behavior. This study highlights early-vigor and earlier senescence over “stay-green” for wheat selection, offering a cost-effective approach shifting the breeding focus from yield maximization to joint yield-stability evaluation, promoting sustainability.
Why it matches plant phenotyping methods高スループットの地上・UAVセンサーによる表現型取得と、機械学習による収量・安定性推定が研究の中心であり、育種選抜に用いる手法を実質的に評価・適用している。
abstractHigh-throughput phenotyping was conducted throughout the season using ground-sensor Red-Green-Blue (RGB) and multispectral (MS) vegetation indices (VIs), along with UAV-derived RGB, MS, and thermal-infrared (TIR) data.
Reproduction assets foundThe authors explicitly state that the datasets and analysis scripts for all analyses (yield/stability modeling, VI extraction, Random Forest workflows) are publicly available in their Zenodo repository (DOI 10.5281/zenodo.17435708), referenced both in the statistical analysis section and the Data Availability statementCode · publicThe datasets and scripts for all the analyses conducted are available in our repository ( https://doi.org/10.5281/zenodo.17435708 ).Open asset ↗zenodo · 10.5281/zenodo.17435708lines:222-237Plant phenotyping relevance match · UnverifiedCrossref · OpenAlex · Europe PMC · checked 5 Sept 2026
Accurate prediction of phenotypes across genotypes and environments is crucial for accelerating crop improvement. Process-based crop growth models (CGMs) can capture complex genotype-by-environment interactions, but their use is limited by labor-intensive genotypic parameter measurements. Here, we developed a faster data assimilation pipeline integrating high-throughput phenotyping (HTP) observations with the SiriusQuality wheat model to efficiently estimate key genotypic parameters and predict genotype performance. Using time-series RGB imagery from a ground-based Phenomobile, we assimilated intercepted photosynthetically active radiation (fIPAR), heading date, and final grain yield to jointly assimilated to calibrate twelve genotypic parameters governing phenology, canopy development, light interception, biomass accumulation, and grain filling. Two data assimilation strategies—a Bayesian DREAM (zs) algorithm and a lookup table (LUT) inversion—were compared through both in silico experiment and eight years of multi-environment field trials of nine durum wheat cultivars. The LUT method demonstrated superior computational efficiency, with prediction accuracy comparable to Bayesian inference on real field data. Multi-year field trials showed that two environments (year / site) were sufficient to reliably characterize genotypic parameters and predict performance across environments. By combining time-series HTP data with ecophysiological modeling, our data assimilation pipeline offers breeders a powerful tool for genotype characterization. It streamlines the process of capturing environmental variance and phenotypic stability, reducing time and effort in crop improvement.
Why it matches plant phenotyping methodsHTP画像を作物成長モデルへ統合するデータ同化パイプラインを開発し、複数アルゴリズムと実圃場データで性能比較・検証しており、表現型取得・推定手法が研究の中心である。
abstractHere, we developed a faster data assimilation pipeline integrating high-throughput phenotyping (HTP) observations with the SiriusQuality wheat model to efficiently estimate key genotypic parameters and predict genotype performance.
Timely, field-scale retrieval of crop biophysical variables is widely regarded as central to data-driven agronomy. In this study, a practical workflow was evaluated in which high-resolution unmanned aerial system (UAS) multispectral imagery was coupled with PROSAIL inversion to map rice canopy traits across three phenological stages. Multispectral and RGB acquisitions were processed, and indices sensitive to chlorophyll, water, and pigment dynamics (e.g., Normalized Difference Red-Edge Index (NDRE), Leaf Chlorophyll Index (LCI), Modified Chlorophyll Absorption Ratio Index (MCARI), Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Structure-Insensitive Pigment Index 2 (SIPI2), Triangular Greenness Index (TGI), and Visible Atmospherically Resistant Index (VARI)) were derived. Leaf and canopy parameters, leaf chlorophyll content (Cab), carotenoids (Car), leaf water content (Cw), dry matter (Cm), mesophyll structure (N), and leaf area index (LAI)—were retrieved via lookup-table (LUT) inversion of PROSAIL. Independent ground measurements were used for validation, and a same-date Sentinel-2 benchmark was performed (subject to cloud constraints). Consistent phenological trajectories were observed: NDRE/LCI and Cab/LAI were found to peak at maximum greenness, while SIPI2 was observed to rise during senescence alongside declining Cab and Cw. Stage-dependent errors were identified in PROSAIL RMSE maps, with the lowest and most homogeneous errors detected at peak canopy. Strong agreement with field data was obtained (R² > 0.98 for most variables at the first date). For Cab, R²/RMSE values of 0.996/1.555, 0.978/2.104, and 0.972/0.2 were recorded across the three dates, respectively. Lower accuracy was produced by Sentinel-2 at field scale (e.g., LAI R²/RMSE ≈ 0.81/0.7; Cab ≈ 0.78/6.5), although useful cross-sensor complementarity was indicated. An operational pathway to within-field mapping of rice biophysics is thereby offered by the “UAS multispectral + PROSAIL” pipeline. The results demonstrate high accuracy at field scale, with phenology-dependent retrievals outperforming Sentinel-2-based estimates, highlighting the potential of UAV-based approaches for precise crop monitoring. Enhanced robustness to phenological change and cloud-related gaps is achieved when red-edge and pigment-ratio indices are fused with physical inversion, and straightforward extensibility to other cereals and management contexts is suggested.
Why it matches plant phenotyping methodsUAVマルチスペクトル画像とPROSAIL逆解析を統合し、イネの生理・構造形質を推定して地上測定で検証するワークフローが研究の中心であり、実質的な植物フェノタイピング手法の適用・評価である。
abstracta practical workflow was evaluated in which high-resolution unmanned aerial system (UAS) multispectral imagery was coupled with PROSAIL inversion to map rice canopy traits across three phenological stages.
Precise monitoring of wheat phenology (BBCH scale) is essential for agricultural optimization, yet UAV-based single-phase monitoring encounters spectral ambiguities where multiple vegetation indices correspond to identical growth stages. A dual-mode framework integrating time-series reconstruction with hybrid deep learning was developed to resolve this limitation. UAV multispectral and digital imagery (333 plots, 2023–2024) enabled reconstruction of daily-resolved vegetation indices, color/texture features, and BBCH stages using Gaussian, PCHIP, and linear fitting to mitigate environmental noise. Synthetic datasets incorporating Gaussian noise (5–100 % relative intensity) simulated field variability. Feature selection was optimized through Competitive Adaptive Reweighted Sampling (CARS) and Variance Inflation Factor (VIF). Hybrid CNN-GRU and CNN-LSTM architectures surpassed standalone networks by resolving spectral ambiguities in single-phase data and leveraging temporal patterns during time-series analysis. Time-series models attained maximum accuracy under noise-free conditions (CNN-GRU: R 2 = 0.90–0.98, RMSE = 3.61–7.65 BBCH units), with accuracy decreasing proportionally to noise intensity. Conversely, single-phase models demonstrated peak performance at 20 % noise intensity (CNN-GRU: R 2 = 0.56–0.70, RMSE = 15.33–17.22 BBCH units), achieving optimal balance between robustness and practicality for real-time farm monitoring. Extreme noise (100 %) distorted feature distributions (7.25–8.73× expansion), validating controlled augmentation. A novel Rate of Phenological Development (RPDW) —quantified as the slope of BBCH progression—was derived to inform breeding programs, while the noise-optimized single-phase approach enables resource-efficient phenology tracking for family farms. This work bridges methodological innovation (adaptive noise strategies, hybrid architectures) with scalable solutions for precision agriculture, advancing UAV-based phenology monitoring in both academic and applied contexts.
Why it matches plant phenotyping methodsUAV画像・時系列再構成・深層学習を統合し、BBCH生育段階を推定するフェノタイピング手法の開発と性能評価が中心である。
abstractA dual-mode framework integrating time-series reconstruction with hybrid deep learning was developed to resolve this limitation.
Plant disease is a significant challenge for agriculture, leading to reduced yield, economic loss, and environmental impact. Leveraging digital photos of plant leaves, convolutional neural networks (CNNs) have emerged as promising tools for disease detection. The methodology involves several steps, including image pre-processing, segmentation, feature extraction using CNNs. Crucially, a diverse dataset comprising images of both healthy and diseased leaves under varying conditions is necessary for training accurate models. Transfer learning, particularly with pre-trained models like ImageNet, can further enhance accuracy, allowing for better performance with fewer training samples. The proposed method demonstrates impressive results, achieving over 95% accuracy, outperforming existing state-of-the-art techniques. This system could serve as a valuable tool for farmers, facilitating timely disease identification and treatment, ultimately leading to increased agricultural yields, reduced financial losses, and the adoption of more sustainable farming practices. Additionally, beyond its practical applications, the proposed system holds promise for advancing sustainable agriculture by promoting environmentally friendly farming methods and contributing to the overall resilience and productivity of agricultural systems.
Why it matches plant phenotyping methods植物葉画像から病害状態を推定するCNN画像処理手法が研究の中心であり、前処理・分割・特徴抽出・データセット構築と精度評価を扱っているため、植物フェノタイピング手法として含める。
abstractLeveraging digital photos of plant leaves, convolutional neural networks (CNNs) have emerged as promising tools for disease detection.
This study investigated the effect of different red:blue (R:B) spectral light ratios on the performance of a multi-task convolutional neural network (CNN) model developed for the automatic classification of four horticultural species and their corresponding phenological stages under controlled artificial lighting conditions. The model was trained and tested using RGB images acquired under five distinct spectral treatments (R:B 1, 3, 5, 7, and 9), and its performance was evaluated using accuracy, precision, recall, F1-score, and Matthews correlation coefficient (MCC). For species classification, the best results were obtained with an R:B 1, achieving an accuracy of 86%, precision of 87%, recall of 85%, F1-score of 85%, and MCC of 0.81. In terms of phenological stage classification, the highest performance was observed at R:B 3 and R:B 5, both yielding 93% accuracy and F1-score, precision and recall above 92%, and an MCC of 0.86. These findings demonstrate that the multi-task CNN model is capable of learning robust and generalizable representations, maintaining high classification performance even under non-optimal spectral conditions. The integration of optimized artificial lighting with intelligent classifiers proves to be a strategic approach for automated monitoring systems in indoor and precision agriculture. Future research should explore the impact of additional spectral components (e.g., green or far-red wavelengths) and the adoption of more advanced neural architectures to further enhance the system’s robustness and scalability.
Why it matches plant phenotyping methodsRGB画像とCNNによる植物の種およびフェノロジー段階の自動分類手法を開発・評価しており、植物状態の取得・推定が研究の中心である。
abstracta multi-task convolutional neural network (CNN) model developed for the automatic classification of four horticultural species and their corresponding phenological stages
Abstract Observer bias and inconsistencies in traditional plant phenotyping methods limit the accuracy and reproducibility of fine-grained plant analysis. To address these limitations, TomatoMAP is introduced as a comprehensive dataset for Solanum lycopersicum . The dataset contains 68,080 RGB images: 3,616 high-resolution macrophotographs (3648 × 5472) with semantic annotations, and 64,464 moderate-resolution images (1080 × 1440) captured from 12 plant poses at four camera elevations. Each image is accompanied by manually annotated bounding boxes for seven regions of interest (leaves, panicle, flower clusters, fruit clusters, axillary shoot, shoot, and whole-plant area) and by labels spanning 50 BBCH classes representing phenologically growth stages. A general cascading structure is proposed. For real-time applicability, models emphasizing the accuracy-efficiency trade-off (MobileNetv3, YOLOv11, and Mask R-CNN) are prioritized and benchmarked against multiple state-of-the-art models. Performance is assessed using accuracy, mAP, inference FPS, and normalized confusion matrices. In a study involving five domain experts, AI models trained on TomatoMAP achieves comparable accuracy levels. Reliability of automated fine-grained phenotyping is supported by Cohen’s Kappa statistics and inter-rater agreement heatmaps.
Why it matches plant phenotyping methodsトマトの多視点画像、器官領域・生育ステージ注釈を備えたデータセットを構築し、画像モデルの精度・効率・専門家一致度をベンチマークしており、植物フェノタイピング手法が中心である。
titleTomato Multi-Angle Multi-Pose Dataset for Fine-Grained Phenotyping.
Reproduction assets foundThe paper's authors publicly release their analysis code (dataset construction scripts for TomatoMAP-Cls/Det and model training/evaluation code) on GitHub. The TomatoMAP phenotype image dataset itself is deposited at e!DAL (10.5447/ipk/2025/14), but no matching URL is present in the allowed list, so only the code assetCode · publicThe scripts for constructing TomatoMAP-Cls and TomatoMAP-Det, as well as the code used for model evaluation, are available at: https://github.com/0YJ/TomatoMAP.Open asset ↗https://github.com/0YJ/TomatoMAPhtml-lines:423-479Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 15 Sept 2026
Abstract In machine learning–driven plant phenotyping, well-annotated image datasets are essential for developing robust models capable of capturing phenological variability across environments. Here, we introduce DeepPhenoTree – Apple Edition, a multi-site, multi-variety RGB image dataset dedicated to the detection of key phenological stages in apple trees. The dataset comprises 48,320 time-stamped RGB images acquired across four European orchards of the Apple REFPOP consortium under contrasting climatic conditions. From this large corpus, a carefully curated subset of 808 representative images was manually annotated. It includes 241,600 expert annotations covering developmental stages from dormant bud to fruit maturity. Images were acquired using a standardized tractor-mounted phenotyping platform equipped with active flash illumination, ensuring consistent lighting conditions across sites and acquisition dates. Phenological structures were annotated following the BBCH scale, with bounding boxes adapted to organ visibility and developmental stage. In addition to the dataset, we provide baseline deep learning experiments to illustrate detection performance and assess model generalization across locations.
Why it matches plant phenotyping methodsリンゴの生育ステージを検出する注釈付き画像データセットと、標準化された撮像プラットフォームおよびベースラインモデルを提供しており、植物フェノタイピング手法・再利用可能データが研究の中心である。
abstractHere, we introduce DeepPhenoTree – Apple Edition, a multi-site, multi-variety RGB image dataset dedicated to the detection of key phenological stages in apple trees.
The objective of this study was to evaluate the accuracy of detecting crop damage caused by wild boar in rapeseed fields using UAV (unmanned aerial vehicle)-derived RGB (red, green and blue) imagery and deep learning segmentation models. The experiments were conducted on rapeseed crops at full maturity shortly before harvest in central-western Poland in 2021. Four convolutional neural network architectures—U-Net (U-shaped network), U-Net++, DeepLabV3+ (deep learning + labelling), and PSPNet (Pyramid Scene Parsing Network)—were benchmarked using two input configurations: RGB imagery alone and RGB combined with the topographic position index (TPI) derived from a digital surface model (DSM). Model performance was assessed using overall accuracy, F1-score (harmonic mean of precision and recall), and Intersection over Union (IoU), with class-specific metrics reported to provide a realistic evaluation of damaged-area detection. For RGB-only data, overall accuracy ranged from 0.957 to 0.972, while damaged-class F1 and IoU reached 0.752 and 0.603, respectively, for the best-performing model (U-Net). When RGB data were supplemented with TPI, overall accuracy and damaged-class metrics changed only slightly, indicating limited benefit from the topographic feature under these field conditions. Non-damaged crop areas were consistently well-classified (F1 > 0.977, IoU > 0.955). These results confirm that UAV-based RGB imagery enables reliable late-season assessment of wildlife-induced crop damage, and that reporting class-specific metrics in spatially independent test sets is essential for realistic performance evaluation.
Why it matches plant phenotyping methodsUAV画像と深層学習セグメンテーションを用いて、ナタネの野生動物被害という植物状態を推定し、複数モデルをベンチマーク・性能評価しているため、フェノタイピング手法が中心である。
titleDeep Learning Segmentation Models for UAV-Based Detection of Crop Damage in Rapeseed Using RGB Imagery
With the increasing cost of labor, smart agriculture has emerged as a key trend for the future of agricultural development. This paper presents an integrated approach for tomato maturity clas-sification and yield estimation using both RGB and multispectral images. The proposed approach consists of three main components: tomato detection, tomato tracking and counting, and maturity classification of tomatoes. YOLOv8 combined with OSNet is first employed to detect tomatoes, while StrongSORT is then adopted to track consistent identities across image sequences. For maturity classification, multiple vegetation indices, including NDVI, GNDVI, and GRRI, are first transformed using principal component analysis, followed by classification using support vector machines, k-nearest neighbors, and neural networks. Tomatoes are categorized into three ma-turity levels: immature, almost mature, and mature. Results demonstrate that the proposed ap-proach can effectively estimate yield of tomatoes at each maturity stage. This capability provides practical support for harvest planning and labor allocation in precision agriculture.
Why it matches plant phenotyping methodsRGB・マルチスペクトル画像からトマトの成熟度と収量を推定する画像解析ワークフローが中心で、果実状態および収量という植物形質を直接評価している。
abstractThe proposed approach consists of three main components: tomato detection, tomato tracking and counting, and maturity classification of tomatoes.
Forest inventories play an essential role in managing and protecting forest resources as well as quantifying carbon stocks. Recent advances in Uncrewed Aerial Vehicles (UAVs) have enhanced capabilities for efficiently monitoring forest dynamics across large geographic areas. RGB cameras are typically preferred for rapid and scalable forest inventory missions owing to three distinct advantages, including low cost, ease of use, and high resolution. However, compared with multispectral or hyperspectral sensors, the limited spectral signals of RGB cameras pose challenges for tree crown detection and classification. The ability of deep learning methods to capture structural and contextual cues from imagery helps alleviate some of the limitations of RGB data. In this study, we propose an Individual Tree Crown (ITC)-based framework leveraging UAV data and advanced deep learning models for inventories of individual trees in dense and natural forests. First, we develop the ITC-based Multi-Task Convolutional Neural Network (ITCMNet), which incorporates multi-scale contexts to simultaneously and accurately identify individual tree crowns, discriminate tree species, and assess tree vitality. Second, structural parameters for each individual crown are extracted to estimate forest carbon storage using species-specific allometric models. Unlike conventional pixel-based methods, our proposed ITCMNet enables precise forest investigations at the ITC level, enhancing both performance and interpretability. We collected a comprehensive dataset consisting of 2456 ultra-high resolution (1.6 cm) UAV RGB images and 27,160 labeled trees across 105 plots distributed in three dense forests and one city park in Germany to evaluate our framework. The ITCMNet demonstrated robust tree crown delineation performance, achieving an F1 score of 0.81. Additionally, our method attained an F1 score in species classification (i.e., 0.54 for angiosperms and 0.76 for gymnosperms) and vitality identification (0.66). Utilizing precise tree parameters, species information, and species-specific allometric models, our carbon storage estimation surpassed current satellite-based carbon products. The carbon stock estimation achieved an R 2 of 0.83 and the carbon storage range in the Bamberg forests is approximately 50 to 110 Mg C/ha. These results show that our proposed framework provides detailed, cost-effective forest inventories, highlighting its potential to support various downstream forestry applications. The dataset and source code are available ( https://www.dlr.de/en/eoc/about-us/remote-sensing-technology-institute/photogrammetry-and-image-analysis/public-datasets/bamforests ; https://github.com/WendyFan52/ITCMNet ).
Why it matches plant phenotyping methodsUAV画像と深層学習により個体樹冠を検出し、樹冠構造、樹種、樹勢などの植物状態を抽出する枠組みを開発・評価しており、植物フェノタイピング手法が中心である。
abstractwe propose an Individual Tree Crown (ITC)-based framework leveraging UAV data and advanced deep learning models for inventories of individual trees in dense and natural forests.
Abstract Plant disease detection and early disease treatment are essential for sustainable crop production. Computer vision for crop science is growing with the advancement in deep learning. The proposed work systematically addresses these issues through three datasets as Plant Village Maize Dataset (D1), Paddy Doctor Dataset (D2), and Sugarcane Leaf Image Dataset (D3) with different classes. The dataset contains 4188, 16225, and 6748 images from dataset sets D1, D2, and D3, respectively. This work has used a Generative Adversarial Network (GAN) to generate a synthetic dataset. Further use data preprocessing, and the data has been resized to 224×224×3. The proposed model use Depth-wise Multiscale Feature Learning ConvoNet (DMFL-ConvoNet) model, which includes the Depth-Wise Convolutional Block (DCB ) block of DMFL-ConvoNet with 3 × 3 and 5 × 5, facilitates the extraction of multiscale plant disease characteristics. Furthermore, it has added 2.5 million parameters. The proposed DMFL-ConvoNet model offers state-of-the-art performance and decreases computational complexity at 33 frames per second, making it ideal for real-time applications. The proposed DMFL-ConvoNet model has been compared with several transfer learning models, including ResNet50V2, InceptionResNetV2, NASNetMobile, EfficientNetV2L, and EfficientNetV2B0 models, and the proposed model has achieved 99.52% data accuracy in the multiple datasets.
Why it matches plant phenotyping methods葉画像から植物病害を分類する深層学習手法の開発・比較が中心であり、植物の病害状態を直接推定するため、植物フェノタイピング手法として採用する。
abstractThe proposed work systematically addresses these issues through three datasets as Plant Village Maize Dataset (D1), Paddy Doctor Dataset (D2), and Sugarcane Leaf Image Dataset (D3) with different classes.
Grapevines ( Vitis vinifera L.) undergo structural and physiological changes throughout the growing season, progressing through distinct phenological stages that require regular monitoring. This dataset consists of high-resolution point cloud data acquired with a stationary terrestrial laser scanner (TLS) to document grapevine development from early leaf development to dormancy. Georeferenced point clouds were generated from 15 TLS scans along two vineyard rows at nine phenological stages. The dataset also includes multispectral and RGB photogrammetric point clouds and orthorectified raster products from an unmanned aerial vehicle survey conducted before harvest. Ground-truth measurements leaf area index, grape production, and pruning wood biomass were collected for each monitored grapevine. As a result, the dataset provides multi-temporal TLS observations that support grapevine structural analysis and development, phenological monitoring, and can be used for the development of AI-based models for precision viticulture.
Why it matches plant phenotyping methodsブドウの生育・構造・フェノロジーを対象とするTLS点群および関連画像データセットであり、植物フェノタイピング用の再利用可能なデータ基盤として中心的です。
titleTLS-grapevine2024: A terrestrial laser scanner point cloud dataset of grapevines at different phenological stages.
Reproduction assets foundThe paper is a Data in Brief article describing the TLS-grapevine2024 dataset itself, publicly deposited on Zenodo with DOI 10.5281/zenodo.16751663. This is a paper-specific, openly available asset containing the TLS point clouds, UAV imagery/rasters, and ground-truth agronomic measurements (LAI, grape production, prunDataset · publicditions: clear sky.
Data source location
Institution: University of Trás-os-Montes e Alto Douro
City/Town/Region: Arroios, Vila Real, Norte
Country: Portugal
Coordinates: 41°17′28.83″N 7°43′17.90″W,
Altitude: 435 m
Data accessibility
Repository name: Zenodo
Data identification number: 10.5281/zenodo.16751663
Direct URL to data: https://doi.org/10.5281/zenodo.16751663
Related research article
None
1.
Value of the Data
•
This dataset covers nine phenological stages of grapevine growth from April 2024 to January 2025, providing multi-temporal terrestrial laser scanner (TLS) observations for structural and phenological analysis.
•
It includes TLS point clouds collected at multiple stages and muOpen asset ↗Zenodo · 10.5281/zenodo.16751663lines:1-50Plant phenotyping relevance match · UnverifiedCrossref · checked 14 Sept 2026
Published23 Feb 2026International Journal of Latest Technology in Engineering Management & Applied ScienceCited by 0 · OpenAlex ↗
Drone-based phenotyping using unmanned aerial vehicles (UAVs) has emerged as a revolutionary approach for high-throughput, precise, and scalable measurement of plant traits critical to crop improvement. This technology integrates advanced imaging sensors—including RGB, multispectral, hyperspectral, and thermal cameras—with sophisticated image processing and artificial intelligence algorithms to non-destructively capture key phenotypic data such as plant height, biomass, canopy temperature, maturity timing, and disease symptoms under natural field conditions. Compared with traditional manual phenotyping and satellite-based remote sensing, UAV phenotyping offers superior spatial and temporal resolution, enabling dynamic monitoring of complex traits such as drought tolerance and disease resistance. Applications span early stress detection, quantitative trait assessment, yield prediction, and accelerating breeding cycles by facilitating objective, rapid selection of superior genotypes across multiple crop species. Despite its transformative potential, challenges remain in standardizing protocols, managing large-scale complex datasets, integrating phenotypic with genomic and environmental data, and providing training resources for widespread adoption. Ongoing advancements in sensor technology, data analytics, open-source tools, and capacity building are poised to cement drone-based phenotyping as a cornerstone technology for sustainable, climate-resilient crop breeding and global food security.
Why it matches plant phenotyping methodsUAV画像・センサーによる植物形質計測を中心に扱う明示的なフェノタイピングレビューであり、手法の応用、技術、課題を総合的に論じている。
abstractDrone-based phenotyping using unmanned aerial vehicles (UAVs) has emerged as a revolutionary approach for high-throughput, precise, and scalable measurement of plant traits critical to crop improvement.
Background Apple scab (AS), caused by the fungal pathogen Venturia inaequalis, is a major disease of apple that manifests as lesions on leaves and fruits. The disease compromises fruit quality and yield, leading to substantial economic losses. Traditional AS assessment relies on visual scoring, which is labor-intensive, subjective, and poorly reproducible. This study proposes a deep learning-based framework to overcome these limitations and to enable an accurate, scalable AS phenotyping approach. Results Deep learning techniques were employed for the object detection and segmentation of AS symptoms in apple fruits. A two-stage fine-tuning process was applied to color images collected under orchard and laboratory conditions using the YOLO foundation model (YOLO11). The model was first trained to detect healthy apple fruits (Model 1) and subsequently refined to segment AS lesions (Model 2) using high-resolution imagery (864 × 864 pixels). Model 1 (Fruit Detection) achieved 0.98 precision, 0.95 recall, and 0.94 mAP50. Model 2 (Lesion Segmentation) achieved 0.64 precision, 0.75 recall, and 0.75 mAP50. The framework supports real-time processing of images and video. Despite challenges such as variable lighting and symptom heterogeneity, the use of high-resolution training data improved the segmentation accuracy (mAP50-95) of fine-scale lesions by over 50% compared to the previous YOLO architecture. Conclusion These results demonstrate that the proposed deep learning-based approach provides a reliable pipeline for automated AS phenotyping. By improving precision and efficiency in both controlled and field environments, the model enhances apple grading assessments and accelerates breeding efforts to identify AS-resistant genotypes. Furthermore, this work establishes a solid foundation for broader applications in real-time plant disease monitoring and future integration of additional apple diseases.
Why it matches plant phenotyping methodsリンゴ果実上の病斑を画像から検出・セグメント化し、植物病害の程度を自動推定する深層学習手法が研究の中心であるため、植物フェノタイピング手法として含める。
abstractThis study proposes a deep learning-based framework to overcome these limitations and to enable an accurate, scalable AS phenotyping approach.
This data article describes a curated RGB-Depth image dataset captured using an Intel RealSense D435 stereo depth camera mounted on an autonomous mobile platform during field deployments at commercial baby broccoli farms in Victoria, Australia. The dataset comprises 1759 paired RGB images (640 × 480 pixels) and corresponding 16-bit depth frames acquired under both daytime (natural sunlight) and night-time (LED illumination) conditions, designed to support research in agricultural computer vision and robotic harvesting. Images were selected from 39,765 raw acquisitions through a reproducible Python curation pipeline applying quality filtering (blur detection, brightness thresholds, corruption detection), perceptual hash-based duplicate removal, and manual review. The final dataset includes 924 daytime and 835 night-time image pairs containing baby broccoli plants at various growth stages. The dataset provides RGB camera intrinsic parameters and pixel-aligned depth maps to enable 3D point cloud reconstruction. Potential applications include developing deep learning models for crop detection and segmentation, validating depth-based size estimation methods, and benchmarking illumination-robust vision systems. All data and curation code are publicly available under a CC BY 4.0 license.
Why it matches plant phenotyping methodsRGB-Depth画像データセットの構築と再現可能なキュレーションを中心とし、作物検出に加えてサイズ推定という植物形質の評価・ベンチマークに利用できるため。
titleA field-acquired RGB-Depth image dataset for computer vision-based baby broccoli detection and size estimation under varying illumination conditions.
Reproduction assets foundThe paper is a data article describing a public Mendeley Data repository containing the authors' field-acquired RGB-D baby broccoli image dataset (1759 image pairs, ground truth diameter annotations, camera intrinsics, and curation/annotation code), directly reproducing the paper's phenotyping measurements and analysisDataset · publicRepository name: Mendeley Data
Data identification number: 10.17632/px5p6zdk6k.3
Direct URL to data: https://data.mendeley.com/datasets/px5p6zdk6k/3Open asset ↗Mendeley Data · 10.17632/px5p6zdk6k.3html-lines:95-155Plant phenotyping relevance match · UnverifiedOpenAlex · checked 14 Sept 2026
Published21 Feb 2026International Journal of Applied Earth Observation and GeoinformationCited by 2 · OpenAlex ↗
• Proposes a UAV remote sensing framework for ratoon sugarcane seedling counting. • Achieves precise segmentation of canopy using unsupervised learning methods. • Integrates multiple feature methods and models; KBest-F with GBR excels (R 2 = 0.7641). • Reveals the core feature contribution mechanism based on SHAP analysis. • Compares deep learning methods and proves the reliability of the work. Accurate and efficient monitoring of seedling emergence is critical for early-stage crop management and yield forecasting in sugarcane production. To meet this practical demand for precise field phenotyping, this study developed a high-throughput phenotyping framework leveraging unmanned aerial vehicle (UAV) remote sensing data and machine learning. This framework addresses the critical agricultural challenges of inefficient manual counting and the need for plot-scale monitoring in sugarcane production by enabling high-throughput sugarcane seedling number prediction through the integration of UAV-acquired RGB and multispectral imagery. Specifically, the sugarcane canopy was accurately segmented from the background using K-means clustering, a step that enabled the extraction of canopy area and the generation of a mask for obtaining canopy-level average features (including vegetation indices and texture features). These features together form a comprehensive feature set. Subsequently, six different feature selection methods were used to optimize the feature set, and eight machine learning models were combined for training and evaluation. The results showed that the combination of Gradient Boosting Regression (GBR) and KBest-F feature selection method yielded the optimal prediction performance, with a coefficient of determination (R 2 ) of 0.7641, a root mean square error (RMSE) of 19.42, and a mean absolute error (MAE) of 15.93. Further analysis identified canopy area, the Normalized Difference Red Edge Index (NDRE), red edge contrast, and green entropy as core predictive features. They collectively contribute over 60% of total feature importance, and their synergistic effects support accurate seedling number estimation. This framework offers an efficient, scalable tool for plot-scale seedling monitoring, with substantial potential for precision field management of high-density crops.
Why it matches plant phenotyping methodsUAV画像、キャノピー segmentation、特徴抽出、機械学習を統合し、サトウキビ苗数という植物状態を圃場スケールで推定する高スループット表現型計測フレームワークが研究の中心である。
abstractthis study developed a high-throughput phenotyping framework leveraging unmanned aerial vehicle (UAV) remote sensing data and machine learning
We present a novel method for accurately counting winter wheat tillers based on RGB images from hand-collected samples. An efficient sample preparation method assembles wheat tillers into bundles from which individual tillers are robustly detected automatically, using classical image analysis. A custom-made user interface ('TillerCounter' program) allows adjusting the automatic detections interactively, which leads to highly accurate tiller counts comparable to the ground truth obtained by manual counting. The key contributions of our work include:1.An efficient method for imaging straw tillers based on bundle assembly.2.An extensive study of the obtained image quality and comparison with the ground truth data from manual counting.3.Demonstration of the approach's high accuracy using correlation analysis (Pearson correlation coefficient R = 0.973 compared to ground truth) and error analysis (root mean squared relative errors below 5 %).
Why it matches plant phenotyping methods小麦分げつ数という植物形態形質を、画像取得・古典的画像解析・専用ソフトウェアで自動推定し、手動計数を基準に精度検証しているため、フェノタイピング手法が中心です。
abstractWe present a novel method for accurately counting winter wheat tillers based on RGB images from hand-collected samples.
Reproduction assets foundThe paper's authors publicly released the TillerCounter GUI source code on GitHub, which implements the Hough-transform-based tiller counting analysis used in this study. The paper also cites original image/count data at Zenodo (10.5281/zenodo.14446564), but no Zenodo URL is present in the allowed URL list, so only theCode · publicThe source code of the TillerCounter GUI is given at https://github.com/agroscope-ch/TillerCounterGui. Original data is given at Zenodo repository: 10.5281/zenodo.14446564Open asset ↗agroscope-ch/TillerCounterGuihtml-lines:163-195