Plant phenotyping methods.

FG-LCNet: A two-stage foreground-guided network for whole-tree litchi counting →

Accurate litchi counting from whole-tree images is essential for yield estimation, orchard management, and plant phenotyping, but remains challenging in real orchards because fruits occur in dense, heavily occluded clusters and vary markedly in scale, illumination, and appearance across ripening stages, particularly when green fruits resemble surrounding foliage. Existing methods have shown promise, but their robustness in complex orchard environments remains limited. To address these challenges, we propose FG-LCNet, a two-stage foreground-guided litchi counting framework. In the first stage, an enhanced fruit-cluster detector improves the localization of small and ambiguous clusters under complex canopy backgrounds. In the second stage, the detected foreground regions are fed into a density-regression network with hybrid attention, while a consistency-based training strategy is introduced to improve robustness to appearance and illumination variations. To support this study, a large-scale litchi counting dataset was established, consisting of 1,126 whole-tree images collected from five orchards and spanning three ripening stages, with approximately 120,000 fruit-level dot annotations and more than 20,000 cluster-level bounding boxes. FG-LCNet achieved the best overall counting performance, with an MAE of 7.44 and an RMSE of 11.01. It showed clear advantages in high-density fruit-cluster scenarios and cross-orchard validation, while maintaining competitive results across orchard-region and maturity-stage subsets. The framework further retained inference efficiency suitable for practical deployment. These results indicate that FG-LCNet provides an effective solution for robust litchi counting and offers potential for other clustered fruit-counting tasks.

Why it matches plant phenotyping methods果実数という植物器官形質を whole-tree 画像から推定する二段階画像解析手法を開発し、データセット構築と交差果樹園検証まで行っており、表現型取得・抽出法が中心である。

abstractwe propose FG-LCNet, a two-stage foreground-guided litchi counting framework.

A lightweight RPB-YOLO11-based detector improves mobile phenotyping of rice panicle blast →

Rice panicle blast detection is an important task in plant disease phenotyping. Field-based detection remains challenging because infected spike regions are often small, sparse, elongated, and affected by overlapping panicles, complex backgrounds, and variable illumination. In this study, we propose RPB-YOLO11, a lightweight YOLO11-based detector designed for rice panicle blast detection. The model uses a Lightweight Ghost Backbone (LGB) to reduce redundant computation. It uses Anisotropic Axial Stripe Attention (A2SA) to represent elongated panicle structures. It also uses Focal Multi-Scale Attention (FMSA) for multi-scale feature refinement and Adaptive Geometric Shape IoU (AGS-IoU) for geometry-aware localization. The model was trained and evaluated on a rice panicle image dataset containing 1,055 training images, 69 validation images, and 169 test images. On the test set, RPB-YOLO11 achieved 76.09% mAP50, 45.44% mAP50-95, 73.98% precision, and 72.75% recall with 6.21 GFLOPs. Compared with the YOLO11n baseline, it improved mAP50, mAP50-95, precision, and recall by 2.73, 2.12, 1.64, and 2.11 percentage points, respectively. An Android-oriented inference application supports local image inference, detection visualization, class counting, and diseased-panicle incidence estimation. These results suggest that RPB-YOLO11 provides a practical approach for image-based rice panicle blast survey.

Why it matches plant phenotyping methodsイネ穂いもちの画像検出モデルを開発・比較検証し、罹病穂率を推定する実用アプリまで構築しており、植物病害状態の画像ベース表現型取得が中心である。

abstractIn this study, we propose RPB-YOLO11, a lightweight YOLO11-based detector designed for rice panicle blast detection.

ADVANCING PLANT DISEASE DETECTION THROUGH STATE-OF-THE-ART DEEP LEARNING MODELS LEVER-AGING EFFICIENTNETV2, VISION TRANSFORMER, AND ENSEMBLE TECHNIQUES →

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

Cognitive UAV-driven agro-surveillance framework for predicting crop stress–induced yield loss using spatio-temporal learning and adaptive irrigation control →

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

PhenoIntel: A Lifecycle-Aligned Multi-Agent Web Application for Verified, Accessible Plant Phenotype Analysis →

Existing conversational plant-phenotyping platforms are difficult for plant scientists to use and lack the reliability scientific research demands: failed analyses are reported as valid measurements rather than flagged as missing, statistical tests run without checking assumptions, predictions carry no uncertainty estimate, and specialised hardware limits accessibility. We present PhenoIntel, a lifecycle-aligned multi-agent web platform that turns the full machine-learning workflow into a reliable, user-friendly phenotyping system. Nine specialised agents divide the analysis into stages, from image collection through model selection, inference, and reporting, rather than handing the whole task to one AI manager. Independent checks separate these stages, and every agent reads from and writes to one shared, fixed-structure record, so an inconsistent output from one stage is caught before it reaches the next. Uncertainty is matched to each model family, conformal prediction, detection-confidence spread, or Monte Carlo Dropout, rather than applied uniformly, and quality thresholds adapt to crop and task instead of one global cutoff. When no suitable model exists, PhenoIntel can propose, validate, and integrate a new one on its own. The model repository spans ten trained models across five crops and four imaging modalities. Classification models reach Macro F1 of 0.78-0.996; object-detection models reach 0.96 mAP@50 with a 54% reduction in counting error over an unoptimised baseline; and a temporal model reaches held-out Macro F1 of 0.7050. PhenoIntel runs in a browser on standard hardware, requiring no GPU, and a 1,200-test automated suite confirms complete pipeline execution. Every result carries calibrated uncertainty, validated statistics, and FAIR-compliant provenance, a combination existing conversational phenotyping tools do not offer.

Why it matches plant phenotyping methods植物フェノタイピングの画像収集から推論・報告までを扱うウェブプラットフォームを開発し、複数モデル、精度、不確実性、検証スイートを評価しており、方法が研究の中心である。

abstractWe present PhenoIntel, a lifecycle-aligned multi-agent web platform that turns the full machine-learning workflow into a reliable, user-friendly phenotyping system.

Time-resolved volatile organic compound profiling enables non-invasive detection of phenological progression in soybean →

Abstract Background and aims Plant volatile organic compounds (VOCs) change dynamically with plant development and in response to environmental conditions. However, their potential as non-invasive indicators of phenological progression remains poorly explored. In this study, we developed a framework integrating automated VOC sampling, time-resolved VOC profiling, and machine-learning analysis for the non-invasive assessment of plant phenology. Using soybean ( Glycine max (L.) Merr.), we investigated whether development-associated temporal variation in VOC emissions could delineate and predict developmental phases. Methods We collected VOCs daily under controlled environmental conditions from 16 to 43 days after sowing, spanning the transition from vegetative to reproductive stages, using an automated sampling system coupled with thermal desorption-gas chromatograph-mass spectrometer (TD- GC-MS). To characterise temporal changes in VOC profiles associated with phenological progression, we analysed the daily VOC data using a multi-step pipeline combining statistical filtering and similarity-based network analysis. We defined VOC-derived developmental phases from similarity patterns in the VOC profiles, then developed and evaluated machine-learning models to predict these phases. Key results Seven VOCs exhibited distinct phase-dependent dynamics, including green leaf volatiles and monoterpenes showing characteristic temporal changes during phenological progression. Network-based clustering of VOC profiles resolved five developmental phases closely aligned with conventional developmental stages. A machine-learning model predicted these phases from the VOC profiles with high predictive accuracy on independent test data, demonstrating that phenological progression could be quantitatively inferred from VOC emission patterns. Conclusions Our findings support VOC profiling as a reliable and non-invasive approach for assessing phenological progression in soybean. By extracting temporally structured VOC signals, this framework captures developmental information that may be difficult to obtain through visual observation alone, particularly after canopy closure. VOC profiling offers a practical tool for monitoring crop developmental dynamics and has broader potential for plant phenotyping and precision crop management.

Why it matches plant phenotyping methods自動VOCサンプリング、時系列VOCプロファイリング、機械学習を統合し、VOCから植物の発育段階を非破壊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中心である。

abstractwe developed a framework integrating automated VOC sampling, time-resolved VOC profiling, and machine-learning analysis for the non-invasive assessment of plant phenology.

An annotated dataset of soybean root nodules for deep learning-based object detection →

Abstract Technological advances have expanded the adoption of digital technologies in agriculture, helping to reduce labour effort, increase profitability, improve crop efficiency and productivity, enhance product quality, mitigate environmental impacts, and promote human health. This context also extends to soybean farming, a sector of major economic importance in Brazil. Most importantly, Brazil has the global leadership in soybean production with biological nitrogen fixation (BNF) replacing chemical fertilisers. The research and evaluation of BNF is limited by manual counting of nodules, a time-consuming procedure. This study presents SoyNodules, designed for the automatic identification of soybean nodules, consisting of a dataset of images. The dataset includes 1,701 images acquired under controlled conditions: 1,662 images of soybean roots with nodules and 39 images of isolated nodules without roots. A total of 49,210 nodule instances are manually annotated with bounding boxes. SoyNodules was designed to promote reuse and interoperability in alignment with the FAIR principles (Findable, Accessible, Interoperable, Reusable) and to support the development, training, and evaluation of computer vision and deep learning methods for precision agriculture.

Why it matches plant phenotyping methods大豆根粒を自動識別する画像データセットであり、手作業計数の代替となる植物器官形質の抽出・評価を支援する方法論的データセット。

abstractThis study presents SoyNodules, designed for the automatic identification of soybean nodules, consisting of a dataset of images.

A Localization-Aware Heterogeneous CNN Ensemble with Neural Meta- Fusion for Plant Disease Classification, with a Component Analysis on Laboratory and Field Images →

Abstract Deep convolutional networks now classify leaf images on curated benchmarks such as PlantVillage with accuracies close to the measurement ceiling of those datasets, which has shifted the open questions away from raw accuracy toward two under-reported issues: which components of a composite pipeline actually cause the result, and whether the components that matter on laboratory images are the same ones that matter on field photographs. We address both with a classification framework evaluated under a protocol that fixes every development decision before the independent test data are read. A Mask R-CNN stage localizes the dominant leaf, the accepted box is expanded by a validation-selected 5% margin and resized to a shared 224 by 224 input, and the same localized image is passed to ResNet50, InceptionV3, and MobileNetV2 together with an extractor that produces eighteen colour and shape descriptors. The forty-five class probabilities and eighteen descriptors form a sixty-three-dimensional input to a neural meta-classifier developed by three repetitions of stratified five-fold cross-validation. On a locked 3,101-image PlantVillage test partition of fifteen classes the framework reached 99.77% accuracy and 99.75% macro F1 with seven misclassifications, and a one-component-at-a-time ablation confirmed that every stage contributed. The same design was then trained and evaluated entirely within a separate thirteen-class PlantDoc field-image dataset, where it reached 90.03% accuracy and 89.40% macro F1. This is a within-PlantDoc experiment and not a controlled-to-field transfer test, so the figure measures how the pipeline behaves on field imagery rather than how a PlantVillage-trained model survives a domain shift. The central finding comes from running the identical ablation on both datasets: on clean images the ensemble breadth and descriptors provide the incremental gains, but under field conditions the ordering changes, and leaf localization and learned fusion become the decisive components. Removing localization cost 2.83 accuracy points and replacing the neural fusion with soft voting cost a further 2.08 points, the two largest effects on PlantDoc. The contribution is a controlled and transparent account of where each component of a localization-aware plant disease classifier earns its place, and of how that ordering shifts between laboratory and field acquisition.

Why it matches plant phenotyping methods葉画像から植物病害状態を推定する分類パイプラインの開発・アブレーション検証が中心であり、単なる病害実験や routine measurement ではない。

titleA Localization-Aware Heterogeneous CNN Ensemble with Neural Meta- Fusion for Plant Disease Classification, with a Component Analysis on Laboratory and Field Images

High-throughput pollen germination phenotyping for assessing heat tolerance in soybean →

Abstract Heat stress causes ultrastructural damage in pollen grains, leading to reduced pollen germination, pollen size and shortened pollen tube length, ultimately lowering seed set and yield. This study presents a high-throughput phenotyping framework that integrates controlled-environment pollen germination assays with deep learning–based object detection for rapid, accurate, and scalable evaluation of reproductive heat tolerance in soybean breeding programs. Sixteen soybean genotypes were grown under controlled environments at optimal (28/18°C; day/night) and high temperature (38/28°C; day/night) regimes during flowering. In vitro pollen germination was quantified using six YOLO (You Only Look Once) object-detection architectures (YOLOv7–YOLOv12) to identify the best-performing model for automated analysis. Among the tested object-detection architectures, YOLOv9 achieved the best overall performance for detecting germinated and non-germinated pollen grains in complex images. High temperature significantly reduced mean pollen germination from an average of 40% under optimal conditions to an average of 21% under heat stress (P < 0.05), with a significant genotype × growth temperature interaction. Invitro incubation temperatures ranging from 10 °C to 45 °C produced a clear thermal response; however, no significant genotype × incubation temperature interaction was detected within either growth temperature regime. Although photosynthetic and physiological traits were measured exploring their relationship with pollen germination, their transient and complex response limited their reliability for predicting reproductive performance. The automated pipeline substantially reduced the time required to evaluate pollen germination. The pipeline processed nearly 5,000 images in approximately one hour, substantially increasing throughput and reducing reliance on manual counting. The findings demonstrate that pollen germination is a promising proxy trait for screening reproductive heat tolerance in soybean. Combining controlled environment phenotyping with YOLO-based object detection enabled efficient, accurate, and scalable pollen analysis, and represents the central methodological advance of this study. YOLOv9 performed best among the tested architectures, although discrepancies from manual counts in some images indicate that additional validation is needed. The weak associations with vegetative physiological traits further support the value of direct pollen-based phenotyping.

Why it matches plant phenotyping methods深層学習による花粉画像解析を中心に、花粉発芽という生殖形質を高速・自動測定するハイスループット表現型解析フレームワークを開発・比較・検証している。

abstractThis study presents a high-throughput phenotyping framework that integrates controlled-environment pollen germination assays with deep learning–based object detection for rapid, accurate, and scalable evaluation of reproductive heat tolerance in soybean breeding programs.

Assessing cotton boll-opening concentration for harvest decision-making via foundation model-enhanced cross-scale phenotyping. →

Boll-opening concentration is critical for mechanical cotton harvesting, yet it is still assessed mainly by manual records and single time-point indicators that miss temporal dynamics. To bridge the lack of a unified workflow linking vision foundation models, multi-temporal boll-opening monitoring, and harvest decision-making, we developed a cross-scale UAV high-throughput phenotyping framework centered on DINO-BollGX. DINO-BollGX couples a DINO v3 backbone, a RetinaNet detection head, and an adaptive refinement-and-suppression module for robust open-boll detection under complex field conditions. Using multi-temporal UAV imagery collected over two years for 383 cultivars, we reconstructed plot-scale time series of open-boll counts, derived dynamic features describing progression and intensity changes, and proposed a Cotton Boll-Opening Temporal Stability Index (CTSI) to quantify boll-opening rhythm and concentration; CTSI was further integrated with a time-based risk function to generate harvest decision curves. Under unified data and training settings, DINO-BollGX achieved precision = 0.91, F1 = 0.88, and AP@0.50 = 0.80, and provided accurate boll-count estimation (R 2 =0.98; MAE=3.10), outperforming YOLOv11, YOLOv12, YOLOv13, and RT-DETR. On an independent cross-year dataset acquired at 5 m altitude, it obtained precision = 0.98 and F1 = 0.87. An internal consistency analysis showed that CTSI had the expected negative association with Window_days (r = −0.83) and positive associations with Max_count (r = 0.78) and the boll-opening efficiency index (r = 0.92), reflecting the co-occurrence of temporal compactness and main-phase opening intensity in the cultivar population. CTSI ranged from −2.72 to 4.69 across cultivars, enabling identification of highly synchronized boll-opening. Harvest decision curves indicated that the relative net income index peaked at day 67 after the first observation and a compact optimal harvest window near the end of monitoring; on a fixed harvest date, Kuche 130292 (CTSI=4.69) produced 486 open bolls versus 182 for Xinluzao 36 (CTSI=0.53) and 90 for Andizhan-60 (CTSI=-2.72). Overall, the framework integrates dynamic boll-opening phenotyping with harvest timing optimization, supporting scalable cultivar screening and mechanization-ready deployment, with potential extension to harvest decision scenarios in other crops.

Why it matches plant phenotyping methodsUAV画像と基盤モデルを用いて綿花の開絮を検出・定量し、時系列表現型指標を構築・検証する方法が研究の中心であるため。

abstractwe developed a cross-scale UAV high-throughput phenotyping framework centered on DINO-BollGX.

Detection of Tomato Leaf Disease in Leaves with Deep Learning MobileNetV2 with Gaussian and Gabor Preprocessing →

.

Why it matches plant phenotyping methodsトマト葉の病害を画像と深層学習で検出する手法が題名上の中心であり、植物の病害状態を観察的に推定するフェノタイピング研究に該当する。

titleDetection of Tomato Leaf Disease in Leaves with Deep Learning MobileNetV2 with Gaussian and Gabor Preprocessing

AI driven multi modal deep learning system for wheat disease detection, yield prediction, and crop health monitoring →

Sustainable wheat farming is challenging. Real-time information on crop health, disease transmission, and anticipated yields is essential for farmers. However, they frequently use slow, expensive, or non-communicative tools. This project develops a workable solution. There is no need for massive server farms because the entire system operates on a single graphics card. It incorporates images of wheat fields, Indian farming notes, greenhouse records, harvest statistics, and NASA meteorological data. Consider them as various “eyes” for crop photo analysis, and we tried several lightweight computer vision models. ConvNeXt-Tiny was slower but could operate on older equipment with 75% accuracy; EfficientNetB0 recognised wheat heads with 92% accuracy; and AgroMark, a hybrid solution that merged photo analysis with agricultural metadata (soil type, rainfall, increased to 87%, etc. Combining picture analysis with attention mechanisms (CBAM) allowed us to anticipate the amount of wheat that a field will yield based on these photo insights, and the results showed that our predictions were accurate, with an R 2 score of 0.97. Additionally, we developed a versatile detector that simultaneously detects disease, stress, head count, and pests. It is adjusted to deal with training data that is unbalanced (some diseases are common, while others are rare). As we packed everything into a 16-GB graphics card, we spent real time determining which strategies smaller training sets, removing weak features, and adjusting loss functions, work. We encounter real-world obstacles along the road, such as photographs from different locations not always match, mislabeled photographs from different locations not always match, mislabeled diseases, and neglected rare pests. Our step-by-step instructions, charts, and code are available.

Why it matches plant phenotyping methods小麦画像から病害・ストレス・穂数・収量などの植物形質・状態を推定するマルチモーダル手法を開発し、複数モデルの精度比較と実装上の検証を行っており、表現型取得・推定が研究の中心である。

abstractThis project develops a workable solution.

A Hybrid Transfer Learning Framework for Corn Crop Detection Using Deep Convolutional Networks →

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.

Attention-driven YOLOv11n-DeiT model for enhanced detection of tomato leaf diseases. →

The tomato plant is considered one of the most important crops in the world, yet it is vulnerable to various diseases that affect crop quality and agricultural productivity. These challenges have driven the need for an efficient and intelligent plant disease detection system. With the development of computer vision and artificial intelligence, this proposed methodology based on deep learning for tomato leaf diseases has been presented. Two public datasets: Taiwan DS with nine classes and Tomato Leaf Diseases Detection Computer Vision Dataset (TLDDCV DS) with seven classes have been used to test this system. This system begins with plant image processing, which includes gamma correction and bilateral filtering, to enhance image quality and clarity while preserving key disease features. Then, a genetic metaheuristic algorithm was used to automatically select the most significant hyperparameters, further optimizing both processing time and accuracy. After that, the tomato leaf disease detection applies the You Only Look Once version 11 Nano (YOLOv11n) model. The YOLOv11n backbone is edited through a Data-efficient Image Transformer (DeiT) to improve the system's capacity for learning global contextual information and long-range dependencies. Experimental results demonstrate that the proposed system outperforms existing methods. It achieved an average mAP@50 of 97.8%, mAP@50-95 of 93.4%, precision of 97.3%, recall of 93.8%, and F1-score of 95.5% on the Taiwan dataset. Additionally, it achieved an average mAP@50 of 87%, mAP@50-95 of 48%, precision of 83.9%, recall of 70.3%, and F1-score of 76.4% on the TLDDCV dataset. These results demonstrate the generalizability and effectiveness of the proposed system in real-world agricultural situations.

Why it matches plant phenotyping methodsトマト葉の病徴を画像から検出・分類する深層学習手法の開発と2データセットでの性能評価が研究の中心であり、植物病害状態のフェノタイピングに該当する。

abstractthe tomato leaf disease detection applies the You Only Look Once version 11 Nano (YOLOv11n) model.

AgriX-SENet: Squeeze-and-Excitation-based deep learning framework for explainable plant disease detection in sustainable agriculture →

Introduction Timely and accurate detection of plant diseases is essential for ensuring global food security and supporting sustainable agriculture. Conventional diagnostic approaches, such as manual inspection and laboratory testing, are often time-consuming, labor-intensive, and impractical for large-scale or remote agricultural environments. Although deep learning models, particularly Convolutional Neural Networks (CNNs), have significantly improved automated plant disease classification, they often lack interpretability and struggle to generalize under diverse field conditions. Methods This study proposes AgriX-SENet, an explainable deep learning framework that integrates Squeeze-and-Excitation (SE) blocks with a DenseNet121 backbone to enhance disease classification performance. The SE blocks recalibrate channel-wise feature responses to emphasize disease-relevant information while suppressing background noise. To improve model transparency, Grad-CAM, SHAP, and LIME were incorporated to provide visual and feature-level explanations of the model’s predictions. The framework was trained and evaluated using the Plant Pathology 2020 dataset containing four classes: healthy, rust, scab, and multiple diseases. Results AgriX-SENet achieved a training accuracy of 97.47% and a validation accuracy of 95.07%, outperforming fourteen state-of-the-art deep learning models. The classification report demonstrated high precision and recall across most disease categories, although the scab class exhibited comparatively lower recall, indicating an opportunity for further improvement. The explainability analyses consistently showed that the model focused on pathologically relevant regions of leaf images, validating the reliability of its predictions. Discussion The proposed AgriX-SENet framework effectively combines high classification performance with model interpretability, addressing a key limitation of existing CNN-based plant disease detection systems. Its ability to provide accurate and explainable predictions makes it a promising solution for scalable agricultural diagnostics. Future work will focus on improving classification performance for challenging disease categories and optimizing the framework for deployment on mobile and edge computing devices to enable real-time field applications.

Why it matches plant phenotyping methods葉画像から植物病害状態を推定する説明可能な深層学習フレームワークを開発・評価しており、植物表現型取得・判定手法が中心である。

abstractThis study proposes AgriX-SENet, an explainable deep learning framework that integrates Squeeze-and-Excitation (SE) blocks with a DenseNet121 backbone to enhance disease classification performance.

Spatial-aware lightweight network for real-time tea disease detection: A coordinate attention-enhanced YOLOv8n approach with path-decoupling strategy. →

The intelligent identification of tea diseases is crucial for ensuring tea quality and reducing economic losses in the tea industry. However, the deployment of deep learning models on edge devices remains challenging due to the conflict between detection accuracy and computational overhead. To address this, we propose CA-YOLOv8n, a lightweight object detection model tailored for tea disease diagnosis. Specifically, we introduce a Path-Decoupling strategy to streamline the network structure and integrate the Coordinate Attention (CA) mechanism to enhance the model's spatial awareness of subtle pathological features. Experimental results demonstrate that the proposed model achieves a mean Average Precision (mAP@50) of 98.89% while reducing the parameter count by 32.6% and FLOPs by 24.1% compared to the baseline YOLOv8n. The model was integrated into a diagnostic platform with an automated reporting interface, demonstrating that real-time tea disease identification is feasible on commodity CPU hardware in resource-constrained agricultural environments.

Why it matches plant phenotyping methods茶葉の病害を画像から検出する軽量深層学習モデルを開発・評価し、植物の病害状態を直接推定する方法が研究の中心である。

titleSpatial-aware lightweight network for real-time tea disease detection: A coordinate attention-enhanced YOLOv8n approach with path-decoupling strategy.

Multi-Crop Leaf Disease Detection using YOLOv12 with Class-Aware Multi-Scale Fusion and Adaptive Attention Modules →

- Enhancing agricultural productivity and attaining sustainable crop management depend on the early and precise identification of leaf disease. Using state-of-the-art technologies in precision agriculture like machine learning (ML) and image processing greatly increases the effectiveness of disease detection and facilitates well-informed decision-making. But conventional manual inspection techniques are still tedious, unpredictable, and prone to errors. In order to overcome these constraints, this research offers YOLOv12-CropNet, an innovative deep learning-based system for multi-crop leaf disease diagnosis in real time. The proposed YOLOv12-CropNet approach makes use of the Convolutional Block Attention Module (CBAM) for adaptive attention, the Content-Aware Reassembly of Features (CARAFE) up-sampling module to preserve fine-grained disease characteristics, the YOLOv12 architecture improved with Ghost Convolution for effective feature extraction, and Involution layers to capture spatially specific patterns. Inspection techniques are still laborious, arbitrary, and prone to mistakes. A substantial set of data of 38 classes of both healthy and sick leaves from a variety of crops, including tomato, potato, apple, grape, corn, mango and sugarcane, was put together for training and evaluation. Experimental results show that YOLOv12-CropNet finds a suitable balance between computational speed and accurate detection. Accuracy, F1-score, recall, and precision are important performance metrics that verify the model's resilience in challenging environmental and visual circumstances. The suggested technique provides a scalable and field-deployable way to assist effective identification of diseases and precision agricultural decision-making. The proposed YOLOv12-CropNet model exhibits better performance than the other evaluated models, attaining a 98.45% peak accuracy, 98.10% precision ,98.20 % sensitivity and a 98.18% F1 score, thereby highlighting its efficacy in multi-crop leaf disease detection.

Why it matches plant phenotyping methods複数作物の葉の病徴を画像から検出・分類する深層学習手法を開発し、データセットと性能評価を伴うため、植物病害状態のフェノタイピング手法が中心である。

abstractthis research offers YOLOv12-CropNet, an innovative deep learning-based system for multi-crop leaf disease diagnosis in real time.

CocoaDeep: A preliminary study of the performance sensitivity to datasets of Faster RCNN, YOLO and transformer networks for cocoa pod detection. →

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.

AI for Precision Fertilizer and Pesticide Application: An Integrated Real-Time Deep Learning and IoT-Driven Field Management System →

Abstract Blanket-rate agrochemical scheduling — a practice wherein the same quantity of fertilizer or pesticide is spread uniformly across an entire field irrespective of spatial or temporal crop need — persists as the dominant farm management paradigm across rural India and large parts of South Asia. This approach generates cascading inefficiencies: excess nitrogen drains into waterways, off-target pesticide deposits devastate pollinators, input costs erode thin profit margins, and wide-scale greenhouse gas release from soil microbial activity accelerates climate change. The study documented here addresses this challenge through a purpose-built, four-layer intelligent field management platform. The platform ingests continuous data from drone-mounted multispectral cameras, in-field IoT soil probes, a wireless weather station, and cloud-sourced Sentinel-2 satellite imagery, then passes these inputs through a cascaded AI inference stack. A fine-tuned YOLOv8-L network performs real-time pest and foliar disease localisation; a ResNet-50 backbone quantifies canopy health across five stress gradients; a two-layer stacked LSTM projects short-horizon yield trajectories; and a Deep Q-Network autonomously plans drone spray routes weighted by field-specific prescription maps. Field validation spanned two consecutive growing seasons (Rabi 2022–23 and Kharif 2023–24) across six georeferenced plots covering 4.8 ha at Baramati, Maharashtra. Outcome metrics recorded during head-to-head comparison with conventional practice included a disease detection score of 95.6% mAP, a 47.3% reduction in total nitrogen applied, a 38.1% decrease in pesticide volume, and a 22.4% uplift in harvested grain weight. Together, these field-verified numbers substantiate the operational readiness of integrated AI precision agriculture for smallholder deployment.

Why it matches plant phenotyping methodsマルチスペクトル画像・深層学習による病害局在化とキャノピー健康状態の定量化を中核機能とする統合プラットフォームであり、植物の病害状態・生育状態を直接推定して現地検証している。

abstractThe platform ingests continuous data from drone-mounted multispectral cameras, in-field IoT soil probes, a wireless weather station, and cloud-sourced Sentinel-2 satellite imagery, then passes these inputs through a cascaded AI inference stack.

MatchPlant: An Open-Source Pipeline for UAV-Based Single-Plant Detection from Undistorted Images with Orthomosaic Projection →

Accurate identification of individual plants from unmanned aerial vehicle (UAV) imagery is essential for high-throughput phenotyping and data-driven decision-making in plant breeding. This study presents MatchPlant, a modular, open-source Python pipeline with a graphical user interface for UAV-based single-plant detection and geospatial trait extraction. The pipeline integrates UAV image processing, user-guided annotation of selected undistorted images, convolutional neural network–based object-detection training, forward projection of bounding boxes onto an orthomosaic, and shapefile generation for spatial phenotypic analysis. This workflow preserves native image geometry during detection while maintaining coordinate traceability from source imagery to georeferenced outputs. Across five independent training runs using early-season maize imagery, MatchPlant achieved source-image-level detection performance of AP@0.5 = 90.3 ± 1.1% and mAP@0.5:0.95 = 43.4 ± 2.9%. Orthomosaic-level evaluation after forward projection showed AP@0.5 = 89.7 ± 0.8% and recall = 93.7 ± 1.2%, demonstrating the workflow’s ability to transfer plant detections into georeferenced outputs. Plant-level traits, including plant height derived from canopy height models and NDVI derived from vegetation index rasters, showed strong agreement with manual annotations ( r = 0.87–0.97). Detection outputs were reused across time points with minimal additional annotation, supporting temporal phenotyping during early growth. The framework was validated using maize imagery from a single site and growing season, where plant separation remained clear. By combining modular design, reproducibility, and coordinate traceability, MatchPlant provides an open-source workflow for UAV-based plant-level analysis, with broader applications requiring validation across additional crops, sensors, growth stages, GSDs, and field conditions.

Why it matches plant phenotyping methodsUAV画像から個体検出と植物形質（草高・NDVI）を抽出する、オープンソースの再利用可能なワークフローを開発・検証しており、植物フェノタイピング手法が中心である。

abstractThis study presents MatchPlant, a modular, open-source Python pipeline with a graphical user interface for UAV-based single-plant detection and geospatial trait extraction.

Optimizing SfM parameters for RGB-only individual-tree detection in loblolly pine (Pinus taeda L.) and mixed pine-hardwood stands →

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

BeeMonitor: Automated IoT video surveillance and an AI-powered video processing system for monitoring the foraging and nesting behavior of cavity-nesting solitary bees →

Solitary bee species that use artificial trap nests are important for agricultural crop production and as indicators of habitat quality. Quantifying cavity-nesting solitary bee foraging and nesting behavior is essential for real-time analysis of population numbers and pollination activity, as well as understanding how environmental conditions shape reproductive success and population dynamics. However, manual observation is labor-intensive, prone to observer bias, and unable to deliver continuous data. Existing automated systems either require individual bee marking or detect presence without resolving nest-tube-level entry and exit events. We developed BeeMonitor, an integrated hardware and computer-vision pipeline that detects nest entry and exit events in cavity-nesting solitary bees from continuous video, using Osmia cornifrons (the horn-faced mason bee) as a model system. A low-cost Raspberry Pi handles solar-powered field recording, while the software combines object detection (YOLOv26), a custom multiple-object tracker (BeeTrack), and a Random Forest classifier trained on trajectory-derived features to distinguish genuine events from incidental detections. Over a 29-day deployment, hardware reliability averaged 97.5% recording coverage. The pipeline achieved 91.3% precision and 87.3% recall (F1 = 0.893), generalizing robustly under leave-one-video-out cross-validation (mean F1 = 0.904). Detected foraging trips correlated strongly with brood cell counts (R2 = 0.849, p < 0.001, n = 19), and a Random Forest model (AUC = 0.820) identified solar radiation as the dominant driver of foraging activity, followed by temperature. BeeMonitor demonstrates that automated computer vision can reliably extract ecologically relevant behavioral data from continuous video, enabling real-time analysis of pollinator behavior and abundance at a temporal and spatial resolution unattainable through manual observation. Its modular design supports adaptation to other species and monitoring contexts.

Why it matches plant phenotyping methods植物ではなく昆虫を対象とするが、映像から採餌・営巣行動を抽出する技術開発として中心的であり、指定スコープの植物表現型ではないため除外。

abstractWe developed BeeMonitor, an integrated hardware and computer-vision pipeline that detects nest entry and exit events in cavity-nesting solitary bees from continuous video

A robust cross-crop disease detection framework based on SIS-YOLOv11 with climate-adaptive mechanisms →

Plant disease detection under complex climatic conditions and cross-crop scenarios remains a critical challenge. To address this, we propose a novel SimAM-Inception-StyleRandomization (SIS)-YOLOv11 algorithm based on YOLOv11n for early/late blight detection on potato and tomato leaves. Our core innovations are: 1) A C3k2-SSI module integrating Style Randomization, Inception architecture, and SimAM attention to enhance cross-crop generalization; 2) A Fusion-InceptionConv module for fine-grained feature extraction under rainfall/haze noise; 3) SPPF-Inception and C2PSA-IS modules to optimize multi-scale feature fusion; 4) DepGraph pruning to reduce 47.82% parameters while improving performance. Experiments show that the pruned SIS-YOLOv11 outperforms YOLOv11n by 3.7% in precision, 6.6% in recall, 5.4% in mAP50, and 7.9% in mAP50-95, and surpasses mainstream models (Faster R-CNN, SSD, etc.). This study provides a robust, lightweight solution for automated cross-crop disease detection in complex agricultural environments.

Why it matches plant phenotyping methodsジャガイモとトマト葉の病害状態を画像から検出する新規アルゴリズムを開発し、性能比較・軽量化まで行っており、植物フェノタイピング手法が中心である。

abstractwe propose a novel SimAM-Inception-StyleRandomization (SIS)-YOLOv11 algorithm based on YOLOv11n for early/late blight detection on potato and tomato leaves.

SPROUT: AI-based seedling emergence PRedictiOn and trait extraction using RGB time-series →

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.

Anisotropic boundary-aware detection for cotton leaf diseases with boundary-decoupled regression and lightweight feature adaptation. →

Detecting cotton leaf diseases in open-field environments is challenging due to cluttered backgrounds, scale variation, and irregular lesion morphology. Conventional detectors rely on isotropic receptive fields and coupled box-regression losses, which limit their ability to localize elongated lesions with poorly defined boundaries. We present an anisotropic boundary-aware detection framework that propagates high-frequency boundary information across four successive pipeline stages. In the backbone, an Anisotropic Morphological Contrast Aggregation module (AMCA) enhances direction-aware representation and lesion-background contrast via re-parameterizable strip convolutions and high-frequency residual extraction. A Dynamic Semantic Boundary Transfer mechanism (DSBT) then captures boundary priors from shallow layers before they are lost to downsampling and injects them into the neck. A Morphological-Spectral Synergistic Feature Pyramid Network (MFS-FPN) preserves these cues during multi-scale fusion through spatial-domain operations compatible with edge hardware. Finally, an Anisotropic Boundary-Decoupled IoU loss (ABD-IoU) independently penalizes each of the four box boundaries and sustains optimization signals in high-IoU regimes via a logarithmic modulation factor. On the self-constructed Complex Cotton Leaf Disease dataset (CCLD; 6,856 images, 6 classes), the method achieves 78.50% mAP@50 and 65.00% mAP@50:95, improving the YOLOv11n baseline by 4.80% and 2.70% with only 2.73 M parameters at 202 FPS. Cross-domain evaluations on PlantDoc and RWD confirm consistent improvements. The framework runs in real time on NVIDIA Jetson edge platforms with INT8 quantization.

Why it matches plant phenotyping methods綿花葉の病斑・病害状態を画像から検出する手法を開発し、複数データセットとベースラインで性能検証しているため、植物表現型取得が中心である。

abstractWe present an anisotropic boundary-aware detection framework that propagates high-frequency boundary information across four successive pipeline stages.

A high-performance detection model ISA-YOLO for eggplant pests and diseases. →

Eggplant (Solanum melongena) is a major cash crop, yet field detection of its pests and diseases remains difficult because disease evidence is simultaneously occluded by foliage, blurred at lesion boundaries, and highly variable in scale. Fruit rot is especially challenging: the waxy epidermis and purple anthocyanin-rich surface reduce chromatic contrast, while infection spreads gradually from water-soaked tissue to necrotic tissue, producing diffuse borders between diseased and healthy regions. In this work, we reframe eggplant disease detection through a context-boundary-scale coupling principle, which states that accurate field detection should jointly model incomplete contextual cues, ambiguous lesion boundaries, and scale-varying symptom morphology rather than optimize these cues independently. ISA-YOLO is proposed as an implementation of this principle on top of YOLOv13 through coordinated context modeling, boundary-aware aggregation, and progressive multi-scale fusion. Experiments on two public datasets show that ISA-YOLO achieves 78.1 and 77.7% mAP at 30.66 and 31.74 FPS, outperforming mainstream detectors in overall trade-off between accuracy and speed. After pruning and quantization, inference speed increases to about 75 FPS while maintaining strong accuracy. These results indicate that the proposed principle provides an effective pathway for accurate and deployable eggplant pest and disease detection in smart agriculture.

Why it matches plant phenotyping methodsナスの病害・害虫を画像から検出するISA-YOLOモデル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フェノタイピング手法に該当する。

titleA high-performance detection model ISA-YOLO for eggplant pests and diseases.

A low-cost "plant-scanner" platform for automated detection of Ustilago maydis infection in maize using deep learning. →

Ustilago maydis is a biotrophic fungus that causes smut disease in maize, leading to tumor formation on aerial parts of the plant. While U. maydis has been a model for plant-fungal interaction studies, no tool has existed to automatically quantify infection symptoms under laboratory conditions for deep learning analysis. To address this, we developed a rotating camera system that captures videos of plants under customized lighting and shutter settings. These videos were used to train machine learning models to distinguish between healthy and infected plants. Two detection approaches have been presented. In the first approach, by employing a naive masking technique and combining classical machine learning classifiers utilizing handcrafted features, the model achieved a reasonable performance, with an Area Under the Curve (AUC) of maximum 0.90 on the Receiver Operating Characteristic in one of the classifiers, showing relatively high sensitivity and specificity. The second approach utilizes pre-trained YOLO11 model for object detection and further classification. The YOLO11-based approach outperforms traditional methods, achieving near-perfect validation accuracy (AUC: 0.99-1.00), demonstrating its superiority for real-time, scalable applications. Our toolset, featuring a cost-efficient and customizable scanning platform with open building-blocks design, provides a valuable resource as a proof-of-concept for unbiased disease symptom detection and scoring, with potential applications in other plant pathology studies. This point enables easy replication and adaptation by other research laboratories which makes the platform robust, scalable and practical beyond our specific application.

Why it matches plant phenotyping methodsトウモロコシの感染症状を画像から自動検出・スコア化する低コスト撮像プラットフォームと機械学習手法の開発が中心であり、植物病害表現型の取得・抽出方法に該当する。

abstractwe developed a rotating camera system that captures videos of plants under customized lighting and shutter settings.

Leaf-by-Leaf Diagnosis: A Custom CNN with Pyramidal Feature Extraction for Plant Disease Classification →

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

SPVD-field: a task-oriented multi-task visual dataset for sweet potato virus disease under real field conditions →

Sweet potato virus disease (SPVD) is one of the most destructive diseases affecting sweet potato production worldwide, causing severe yield losses and posing a significant threat to food security. Vision-based intelligent diagnosis has emerged as a promising solution for large-scale SPVD monitoring due to its low cost and scalability. However, existing publicly available datasets for SPVD are extremely limited and typically focus on a single task, such as disease classification or lesion segmentation, under constrained imaging conditions. This lack of comprehensive, task-oriented datasets significantly restricts the development, evaluation, and fair comparison of advanced computer vision methods for SPVD analysis. In this study, we present SPVD-Field, a task-oriented multi-task visual dataset suite composed of two independently collected sub-datasets optimized for different computer vision tasks. Rather than constructing a single homogeneous dataset, SPVD-Field is deliberately organized into two complementary task-oriented sub-datasets: SPVD-DET, designed for disease detection with bounding-box annotations, and SPVD-SEG, designed for fine-grained lesion segmentation with pixel-level masks. The two sub-datasets were independently collected using different acquisition protocols optimized for their respective tasks, while sharing a unified semantic definition of SPVD symptoms, crop growth stages, and field environments. SPVD-Field captures substantial real-world variability in imaging scale, viewpoint, illumination, background complexity, and symptom manifestation, reflecting the inherent challenges of fieldbased disease diagnosis. We provide detailed documentation of data acquisition, annotation strategies, and quality control procedures, along with baseline benchmark results for both detection and segmentation tasks to demonstrate the usability and difficulty of the dataset. By offering a structured dataset suite rather than a single-task collection, SPVD-Field aims to support diverse research directions, including detection, segmentation, multi-task learning, and disease severity analysis, and to facilitate reproducible and comparable research in SPVD-related plant phenotyping.

Why it matches plant phenotyping methodsサツマイモの病徴を対象とする画像データセットで、検出・病斑セグメンテーション、データ取得・アノテーション・品質管理、ベンチマークを中心的に提供しており、植物病害状態の画像フェノタイピング手法・データ基盤に該当する。

abstractIn this study, we present SPVD-Field, a task-oriented multi-task visual dataset suite composed of two independently collected sub-datasets optimized for different computer vision tasks.

Rapid detection and quantification of sweet potato storage roots using ground penetrating radar. →

Sweet potato is a nutritionally valuable crop that contributes to food security, owing to its storage roots rich in starch, sugars, and antioxidants, while requiring minimal cultivation inputs. Estimating its yield based on visible above-ground traits remains challenging due to weak and inconsistent correlations between shoot biomass and storage root development. Therefore, direct assessment of underground biomass is essential. In this study, we demonstrate the field application of ground penetrating radar (GPR) for non-destructive detection and yield estimation of sweet potato. GPR is a geophysical technique that typically transmits ultra high frequency radio waves into the soil and records reflections from subsurface objects. Electromagnetic wave simulations within the soil-root system revealed GPR signals that strongly correlate with root length, forming the basis for yield quantification. We developed an image-processing pipeline comprising static correction, gain adjustment, noise filtering, and hyperbola segmentation via the Hough transform to enable semi-automated storage root detection from GPR data. By integrating detection and quantification approaches, a linear regression model predicting sweet potato yield from GPR signals achieved moderate accuracy ( R 2 = 0.567, normalized RMSE 0.190). We established a non-destructive and low-labor approach for monitoring root systems, providing a foundation for rapid, scalable, and field-ready yield estimation in sweet potato and other root and tuber crops.

Why it matches plant phenotyping methodsGPRによる地下貯蔵根の検出・定量化と収量推定を中心に、信号処理および画像処理パイプラインを開発・評価しているため、植物フェノタイピング手法として収載する。

abstractWe developed an image-processing pipeline comprising static correction, gain adjustment, noise filtering, and hyperbola segmentation via the Hough transform to enable semi-automated storage root detection from GPR data.

Lightweight real-time detectors of apple-leaf diseases operating on embedded devices. →

Agricultural leaf disease detection is crucial for early intervention and yield protection in precision agriculture. Among representative economic crops, such as apples, leaf lesions are typically small and appear in complex backgrounds, making accurate detection performed on resource-constrained embedded devices challenging. To address this, we propose a lightweight small-object detection models, namely the dynamic Differential Compensation Lightweight-YOLO (DCL-YOLO) model and its pruned version (DCL-YOLO-P), based on YOLO11n. A novel Dual-Aspect Feature Complementary Mapping (DAFCM) module type is embedded in their backbone to recover lost semantic and spatial information, while the original YOLO11n's neck is replaced by an Efficient Enhanced Cross-Scale Feature Fusion (EE-CSFF) module, which incorporates Gated Differential Convolutional Fusion (GDCF) modules to strengthen cross-scale information flow and small-object representation. Experimental results obtained on the ALDSOD dataset show that, compared with the YOLO11n baseline, DCL-YOLO improves recall from 81.9% to 84.6%, mAP50 from 86.8% to 88.4%, and mAP50:95 from 47.0% to 47.8%, while also reducing the parameter count from 2.58 M to 1.91 M and Giga Floating-Point Operations (GFLOPs) from 6.3 to 5.5. After applying Layer-Adaptive Magnitude-based Pruning (LAMP), the parameter count and GFLOPs are further reduced to 0.75 M and 2.7, respectively, with mAP50 and mAP50:95 still exceeding the baseline by 1.2 and 0.5 percentage points, respectively. When deployed on an embedded device, the pruned model achieved 15.2 FPS and 139 msec per image, confirming its applicability in real-time scenarios. Furthermore, cross-domain validation, performed on the Global Wheat Head Detection (GWHD) dataset, indicates the stable generalization capabilities of the proposed models across environmental domain shifts. The DCL-YOLO's source code is publicly available at: https://github.com/q123-code/dcl-yolo.

Why it matches plant phenotyping methodsリンゴ葉の病斑を画像から検出する軽量モデルを開発し、データセットで性能比較・クロスドメイン検証・組込み機器での実装評価を行っており、植物の病害状態推定手法が研究の中心です。

abstractTo address this, we propose a lightweight small-object detection models, namely the dynamic Differential Compensation Lightweight-YOLO (DCL-YOLO) model and its pruned version (DCL-YOLO-P), based on YOLO11n.

BerryBox: An affordable computer vision system for postharvest phenotyping of cranberry and other small fruits →

Abstract Fruit size, shape, color, and percent fruit rot are important quality traits for breeding cranberry ( Vaccinium macrocarpon Ait.). Image analysis can be used to measure these traits, but affordable hardware for standardized image capture and integrated user‐friendly software pipelines are lacking. Additionally, no image‐based method exists to estimate percent fruit rot, an otherwise tediously and subjectively measured trait. We created the BerryBox, a simple and inexpensive lightbox, camera mount, and accompanying software pipeline to standardize the capture and analysis of postharvest fruit images. Trained deep neural network models were highly accurate for segmenting sound fruit (F1 score: 99.4%) and detecting rotten fruit (F1 score: 98.5%). We applied the BerryBox to images of cranberries harvested across 3 years from a 156‐clone breeding population. Narrow‐sense heritability estimates of image‐based fruit color, shape, size, and percent fruit rot ranged from 0.37 to 0.95. Random subsampling showed that 25–30 berries per genotype were sufficient to describe the variation in the full dataset. We demonstrated the utility of BerryBox traits in a small‐scale genetic linkage mapping analysis, detecting significant marker–trait associations that coincided with those of traditionally measured traits. The BerryBox software was able to accurately segment fruit from images of blueberries without model retraining, showing its applicability to other similarly shaped fruits. The software pipeline and BerryBox materials and assembly instructions are publicly available for others to adopt for low‐cost image‐based phenotyping.

Why it matches plant phenotyping methodsクランベリー等の果実形質と腐敗率を画像から抽出する低コスト撮像装置・ソフトウェアパイプラインを開発し、精度検証と他果実への適用性評価を行った、中心的な植物フェノタイピング手法研究である。

abstractWe created the BerryBox, a simple and inexpensive lightbox, camera mount, and accompanying software pipeline to standardize the capture and analysis of postharvest fruit images.

An Embedded AI System for Automated Crop irrigation and pest Monitoring →

Modern agriculture is rapidly adopting Artificial Intelligence (AI) and Internet of Things (IoT) technologies to improve crop monitoring and decision-making. Many existing systems focus either on water stress detection or pest detection separately. The proposed system integrates both functions into a single platform. It uses a camera module and environmental sensors connected to a Raspberry Pi (5/4) as the main controller. A Convolutional Neural Network (CNN) model processes leaf images captured by the AI camera, while a soil moisture sensor supports water stress analysis. The system classifies crops into three categories: healthy, water-stressed, and pest-infected. Based on the output, it provides real-time recommendations for irrigation and pesticide application. This reduces manual inspection, prevents unnecessary chemical usage, saves water, and improves crop productivity.

Why it matches plant phenotyping methods植物の葉画像と土壌水分センサーを用いて、健康・水ストレス・害虫感染という植物の状態を自動分類する統合センシング基盤を開発しており、表現型取得・判定が中心的です。

abstractThe proposed system integrates both functions into a single platform.

Paddy leaf disease detection and classification using improved Gorilla Troops optimized YOLO-V8 network. →

Paddy leaf disease (PLD) detection has grown more difficult, yet early detection might prevent significant losses due to decreased crop yield. However, existing models struggle to accurately classify diseases under difficult circumstances like intricate backgrounds, fluctuating lighting, and overlapping leaves. Additionally, existing models do not incorporate efficient optimization strategies, leading to suboptimal accuracy and poor generalization on unseen data. To address these challenges, a novel deep learning-based YOLO-LEAFNET method for PLD detection utilizing IGT-YOLO, integrating the YOLOv8 disease detection with the Improved Gorilla Troops (IGT) optimization. The input paddy leaf images are pre-processed using Bilateral Contrast Limited Adaptive Histogram Equalization (B-CLAHE) to enhance image quality and improve local contrast while preserving disease boundaries. YOLOv8 model is utilized to detect and classify paddy leaf diseases by accurately localizing affected regions with bounding boxes. Then, the IGT algorithm boosts the disease detection accuracy by optimizing YOLOv8 through effective hyperparameter tuning. The proposed YOLO-LEAFNET method effectiveness was evaluated using recall, F1 score, specificity, accuracy, and precision. B-CLAHE enhanced noise-free images improve contrast and detection accuracy, while the IGT-YOLO model ensures scalable, efficient early diagnosis of paddy leaf diseases with 99.07% accuracy. The YOLO-LEAFNET enhanced the total accuracy by 3.21%, 5.25%, and 1.98% related to CNN, DeepRice, and FasterR-CNN, respectively.

Why it matches plant phenotyping methodsイネ葉の病害状態を画像から検出・分類するYOLOベース手法を提案し、前処理・最適化・性能評価を中心に扱っているため、植物フェノタイピング手法として該当する。

abstractTo address these challenges, a novel deep learning-based YOLO-LEAFNET method for PLD detection utilizing IGT-YOLO, integrating the YOLOv8 disease detection with the Improved Gorilla Troops (IGT) optimization.

TriAttnNet based deep learning model for automated cotton pest detection and disease classification. →

This paper presents a deep learning model to detect cotton plant pests and classify diseases, which must overcome limited datasets, class imbalance, and feature redundancy. At the preprocessing phase, the Gaussian blur filtering and Contrast Limited Adaptive Histogram Equalization (CLAHE) are used to sharpen images by improving their clarity and contrast. To increase and diversify the data, Spa-GAN-based data augmentation is used to produce realistic synthetic samples. To obtain an accurate Region of Interest (RoI), an Attention-Guided Multi-Scale Residual U-Net (AGMS-U-Net) is considered to segment local and global structural information. The proposed framework makes three major contributions: (i) a new attention-based feature extractor, TriAttnNet, that incorporates spatial, channel, and contextual attention to represent diseases on a fine-grained level (ii) a new optimization strategy, Hybrid Mongoose Ray Chaotic Optimization (HMRCO), which includes chaotic strategies to better tune the parameters and explore the feature space and (iii) classification layer with focal loss for final decision. Experimental analyses prove that the suggested method is much more effective than the current state-of-the-art models, providing a powerful and understandable solution to precision agriculture and sustainable cotton crop health monitoring. Experimental results show that TriAttnNet achieves 98.66% accuracy, 98.71% recall, and 98.81% F1-score, which is better than the state-of-the-art algorithms, such as EfficientNetB1-CBAM (96.38%) and BERT-ResNet-PSO (95.69%). The proposed system is computationally feasible and interpretable, and it is interpretable to provide a practical solution to precision agriculture and sustainable monitoring of the health of cotton crops.

Why it matches plant phenotyping methods綿花植物の画像から病害を分類する深層学習手法を開発・評価しており、植物の病害状態を直接推定する方法が研究の中心である。

abstractThis paper presents a deep learning model to detect cotton plant pests and classify diseases

Detection of Maize Leaf Diseases Based on MAC Deep Learning →

Abstract Maize is a globally important food crop, and its yield and quality are vulnerable to various leaf diseases. To address issues such as blurred edges of disease spots and difficulty in small target detection, this study proposes the MAC-YOLO11 model improved on the basis of YOLOv11m. The model introduces the MSPP module to enhance the extraction capability of edge and directional features, and incorporates the spatial position attention module CDSA to strengthen the modeling capability for differences in lesion morphology, texture, and spatial distribution. We designed the APC module to expand the effective receptive field at a low parameter cost through asymmetric convolution branches. The dataset covers 11 categories: northern leaf blight, brown spot, common rust, smut, downy mildew, fall armyworm larval damage, gray leaf spot, maize streak virus disease, adult corn borer damage, corn borer larval damage, and healthy maize leaves. Results show that the MAC deep learning model achieves mAP50 and mAP50:95 of 94.1% and 83.5%, respectively, with overall performance superior to YOLOv11m and other mainstream models. This study provides a technical solution for intelligent identification of maize diseases and holds significant value for disease monitoring and precise control in smart agriculture.

Why it matches plant phenotyping methodsトウモロコシ葉の病斑形態・テクスチャ・空間分布を画像から識別する深層学習モデルを開発・評価しており、植物病害状態の画像ベース表現型計測が中心である。

abstractthis study proposes the MAC-YOLO11 model improved on the basis of YOLOv11m.

Rapid Classification and Deep Learning-Based Development Estimation of the Seeds of Helianthus annuus . →

Manually counting sunflower seeds on capitula is labor-intensive, requiring approximately one person-hour per head, and can be inconsistent for densely packed heads. Existing phenotyping approaches often depend on laboratory-based equipment, limiting their accessibility. In this study, we developed a benchtop image-based pipeline for rapid, non-destructive estimation of developed and aborted seeds on intact dried sunflower heads. A dataset of 1093 sunflower capitula was imaged under fixed indoor lighting, and individual seeds were annotated as developed or aborted. A YOLOv8m one-stage object detector was trained and evaluated using a counting-focused protocol, in which a single confidence threshold was selected on the validation set and then applied unchanged to an independent test set of 109 images. The baseline model was compared with recent YOLO variants and different augmentation strategies. On the test set, the model achieved a mean absolute count error of 61.3 seeds per image, a mean relative error of 12.0%, and an mAP50 of 0.18 at the locked confidence threshold of 0.15. Only 13.8% of test images had relative errors below 2%. Larger YOLO models and augmentation variants did not improve performance. These findings show that the proposed system provides approximate, non-destructive seed-count estimation under controlled imaging conditions, while highlighting the need for improved localization in dense regions and domain adaptation for fresh heads or field conditions. The annotated dataset and trained model weights are made available to support reproducible research.

Why it matches plant phenotyping methodsヒマワリ頭花の発達・不稔種子数という植物形質を、画像取得とYOLOによる推定パイプラインで定量化する手法を開発・評価しており、方法が研究の中心である。

abstractwe developed a benchtop image-based pipeline for rapid, non-destructive estimation of developed and aborted seeds on intact dried sunflower heads.

An efficient YOLO-based framework for multi-class plant disease detection →

Plant health plays a critical role in agriculture, climate balance, and economic stability. However, plant diseases caused by bacteria, fungi, and viruses can significantly reduce crop productivity if not detected early. Traditional manual inspection methods are time-consuming, labor-intensive, and prone to human error, especially in large-scale farming. To address these challenges, this study proposes an automated and accurate plant disease detection system using deep learning-based object detection models for early disease diagnosis in agriculture. A publicly available dataset containing 38 different plant leaf diseases annotated in You Only Look Once (YOLO) format is used, along with a standardized preprocessing pipeline to ensure data quality and consistency. Three modern architectures: YOLOv8, YOLOv11, and YOLOv26 were trained and evaluated under identical conditions using the Ultralytics framework on Google Colab. Experimental results show that YOLOv11 achieves the highest accuracy in terms of precision, recall, and mean Average Precision (mAP), while YOLOv8 provides the fastest inference speed with lower computational complexity. Based on the results, the study concludes that YOLO-based models show great potential for plant disease detection, with YOLOv11 offering superior detection accuracy among the evaluated models. The practical implications of these findings lie in the potential for precision agriculture to monitor diseases in real-time, minimize crop losses, and aid in timely decision-making for farmers and agricultural stakeholders.

Why it matches plant phenotyping methods植物葉の病害状態を画像から検出するYOLOベース手法を提案し、複数モデルを同一条件で評価しているため、病害表現型の取得・抽出が中心である。

abstractthis study proposes an automated and accurate plant disease detection system using deep learning-based object detection models

Multi-scale closed-loop tuning via spatial frequency collaborative sensitivity for rice leaf disease detection. →

Rice is a fundamental food source for more than half of the global population, making stable yields and quality improvements vital for food security and sustainable agricultural development. Early infections of rice leaf diseases often exhibit subtle symptoms, while conventional control methods based on empirical judgment and routine pesticide application result in both yield losses and environmental pollution. A Multi-scale closed-loop tuning via spatial frequency collaborative sensitivity (MCCA-YOLO) model has been proposed in this paper with a multiscale closed-loop tuning and spatial frequency collaborative attention mechanism for the early detection and classification of rice crop diseases. MCCA-YOLO incorporates a closed-loop tuning compound network architecture that combines a dual-backbone feature extractor with a spatial frequency enhancement module to achieve system self-verification feedback, reducing transmission errors and enhancing the texture features of leaves. The framework implements a cross-scale weighted fusion and a deformable spatial hybrid attention enhanced bidirectional feature pyramid fusion network for dynamic feature adaptation, effectively accommodating the complex morphology of rice leaf lesions. By conducting comprehensive ablation studies and comparative experiments with existing techniques on the rice plant diseases v8 dataset, the proposed approach achieves a mean average precision (mAP) of 92.2%, outperforming well-established methods, while delivering superior precision (0.915) and recall (0.900). Extensive empirical validation of additional v9 and Rice Leaf Spot Disease (RLSD) datasets for rice plant diseases further demonstrates the model's outstanding performance.

Why it matches plant phenotyping methodsイネ葉の病徴を画像から検出・分類するYOLOベース手法を開発し、アブレーション、比較実験、複数データセットで性能検証しており、植物病害表現型の取得手法が中心である。

abstractA Multi-scale closed-loop tuning via spatial frequency collaborative sensitivity (MCCA-YOLO) model has been proposed in this paper with a multiscale closed-loop tuning and spatial frequency collaborative attention mechanism for the early detection and classification of rice crop diseases.

Explainable CNN framework for accurate crop disease detection using plant leaf images →

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.

Hybrid multimodal learning framework for crop disease detection, adaptive treatment, and price forecasting →

Crop diseases play a significant role in food production globally; therefore, there is an urgent need to develop quick and accurate diagnostic techniques that are more effective than manual inspection methods. The proposed hybrid multimodal learning framework in this research provides a solution that integrates adaptive therapy suggestion, market price prediction, and image-based disease detection. This study also proposes a framework for pesticide recommendation and the treatment of plants. This study experiment on tomato and cotton crop leaf data for disease detection. Experimental results on a tomato crop disease detection dataset show that the proposed model shows high performance. EfficientNetB0 provides more stability and generalization capabilities in different scenarios compared to other models, such as YOLOv8, ResNet50, and a custom CNN model. The use of a knowledge-based decision support system provides sustainable pesticide recommendations based on environmental and symptom-specific parameters. Forecasting of pesticide prices through LSTM methods yields forecasts within 3.2% and 4.1% MAE, enabling improved decision-making by providing instant points of reference for potential price movements. Research uses SHAP and LIME to provide explainability to users, thus improving user buy-in through transparency. Overall, this modular system provides a data-driven decision-making model to improve the efficiency of managing crops.

Why it matches plant phenotyping methods植物葉画像から病害状態を推定する画像ベース手法を、複数モデルで比較評価しており、植物病害フェノタイピングがシステムの主要構成要素です。価格予測や農薬推薦も含みますが、病害検出の技術評価が明示されています。

abstractThe proposed hybrid multimodal learning framework in this research provides a solution that integrates adaptive therapy suggestion, market price prediction, and image-based disease detection.

Wheat spike and spikelet detection and counting from high-resolution digital imagery using YOLO with Oriented Bounding Boxes. →

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.

RubberFormer: a transformer-based detection benchmark for rubber tree powdery mildew. →

Introduction Rubber tree powdery mildew is a major foliar disease that threatens the yield and quality of natural rubber. Its lesions are typically small, irregular, and embedded in complex backgrounds, making accurate automated detection difficult. Methods To address this challenge, we propose RubberFormer, an end-to-end detection framework based on a refined Transformer architecture for detecting small powdery mildew lesions in complex scenarios. RubberFormer adopts MobileNetV4 as a lightweight backbone, introduces the Hierarchical Attention with Local-global Optimization (HALO) module for multiscale local-global feature fusion, incorporates the Unified Cross-Attention Network (UCAN) to enhance multidimensional feature interaction, and applies Normalized Wasserstein Distance (NWD) Loss to improve small-object localization. Results Extensive experiments were conducted on PM-Dataset-Plus, which contains 9,765 images, and PD-40, a large-scale plant disease dataset containing 80,369 images across 40 disease categories and 8 crops. RubberFormer achieved superior detection accuracy and generalization performance compared with existing methods, while maintaining computational efficiency suitable for practical agricultural monitoring. Discussion These results demonstrate that RubberFormer is effective for detecting small and irregular rubber tree powdery mildew lesions under complex conditions. The framework has practical value for rubber tree disease monitoring and provides a transferable design strategy for agricultural vision tasks involving small objects and complex backgrounds.

Why it matches plant phenotyping methodsゴム樹の病斑という植物の病害状態を画像から検出するTransformer手法を開発し、複数データセットで性能検証しており、植物表現型取得が中心である。

abstractwe propose RubberFormer, an end-to-end detection framework based on a refined Transformer architecture for detecting small powdery mildew lesions in complex scenarios.

Lightweight Visual Detection Framework for Complex Background Grape Leaf Disease Identification →

Abstract Accurate crop disease detection supports precision agriculture, but field-deployable identification remains hindered by complex backgrounds, varying illumination, and heavy deep learning models. This work presents a lightweight visual detection approach for grape leaf diseases under unconstrained field conditions. Built on the YOLO11n backbone, the method integrates three customized modules: C3k2-UltraLightBlock for efficient feature representation, LeafRepFusionStem for low-level feature enhancement, and RCSA-HSFPN for refined multi-scale fusion with residual channel-spatial attention. A dedicated dataset with complex backgrounds is constructed via augmentation and background replacement. Experiments show the model achieves 92.0% precision, 92.9% recall, and 93.0% mAP@0.5, with only 2.9 GFLOPs and 1.73 M parameters, representing 54.7% and 33.2% reductions over the baseline. Heatmap visualization confirms improved lesion focusing and background suppression, while cross-crop tests validate strong generalization. This framework provides an efficient solution for real-time, edge-deployable plant disease monitoring, balancing accuracy and computational efficiency for practical agricultural visual computing applications.The implementation code for this study is available at:https://github.com/aitizc/Lightweight-Visual-Detection-Framework-for-Complex-Background-Grape-Leaf-Disease-Identification.git

Why it matches plant phenotyping methodsブドウ葉の病斑・病害状態を画像から推定する軽量な視覚検出手法を開発・評価しており、植物病害表現型の取得が中心である。

abstractThis work presents a lightweight visual detection approach for grape leaf diseases under unconstrained field conditions.

A Lightweight Convolutional Neural Network with Neighbourhood Attention and a 100- Category Dataset for Plant Disease Detection →

Plant disease detection is vital for agricultural sustainability and food security. While Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs) have achieved high accuracy in this domain, CNNs often require millions of parameters and substantial computation. ViTs suffer from the quadratic time and space complexity of self-attention (SA), limiting their use on resource-constrained devices. Although SA is capable of modelling long-range dependencies when symptoms are dispersed, many plant diseases exhibit small, localized lesions or texture changes; therefore, Neighborhood Attention (NA) offers a more efficient and targeted alternative by focusing on nearby regions rather than the entire image. This work proposes a custom Localized NA block implemented in TensorFlow/Keras that operates directly on CNN feature maps, bypassing patch embedding and transformer modules. A lightweight CNN is then developed by combining depth-wise separable convolutions with the proposed localized NA block. In addition, a 100-category plant disease dataset covering 16 crops is presented. The dataset is curated, class-balanced, and made publicly available to support reproducibility and encourage further research. The proposed 9-layer CNN, with just 1.7M parameters and a size of 6.74 MB, achieved a favorable balance between accuracy, model size, and computational efficiency, compared with MobileNetV1, MobileNetV2, DenseNet121, InceptionV3, MobileViT-XXS, and EfficientViT-M0, achieving 98.97%± 0.33% accuracy on PlantVillage and 93.36%± 0.28% on the proposed dataset. The ablation study showed that the NA block improved test accuracy by approximately 2–3%, while Grad-CAM visualizations indicated more precise targeting of diseased areas in the leaf image.

Why it matches plant phenotyping methods植物葉画像から病徴を推定する軽量CNNと注意機構を開発し、複数データセットで比較評価・アブレーションを行い、さらに100カテゴリの公開データセットを提示しているため、植物フェノタイピング手法が中心である。

abstractThis work proposes a custom Localized NA block implemented in TensorFlow/Keras that operates directly on CNN feature maps

Research on an improved RT-DETR-based model for rice disease detection. →

Monitoring and precisely localizing rice diseases is essential for agricultural productivity and food security. Existing detection methods face challenges such as high computational complexity, semantic information loss, difficulty detecting small targets, and limited robustness. To address these issues, this study proposes ECL-RTDETR, an enhanced RT-DETR-based rice disease detection model. First, a lightweight EfficientViT backbone is employed for feature extraction, incorporating a streamlined multi-head self-attention module to improve inference speed, reduce computational cost, and strengthen local feature extraction. Second, the CARAFE upsampling operator is introduced to better preserve detailed feature information without added computational burden, enhancing fine-grained representation. Finally, standard convolution in the neck network is replaced with LDConv (lightweight dynamic convolution) to enable adaptive feature learning under complex conditions, addressing variations caused by illumination, occlusion, and disease diversity. Experimental results show that ECL-RTDETR improves mAP@0.5 by 0.7%, increases detection speed by 22.2 FPS, and reduces computational cost by 81.8 GFLOPs and parameters by 22.12M compared with the baseline RT-DETR. Overall, ECL-RTDETR delivers superior accuracy, speed, and efficiency, offering a robust solution for intelligent rice disease detection and localization, and advancing smart agriculture and sustainable food security.

Why it matches plant phenotyping methodsイネ病害の検出・局在化を目的とする画像解析モデルを開発し、精度・速度・計算量を実験的に比較検証しており、植物の病害状態を推定する方法が研究の中心である。

titleResearch on an improved RT-DETR-based model for rice disease detection.

Transformer-based hybrid classification for plant leaf disease detection using vision transformer, principal component analysis, and support vector machine →

Plant diseases remain a critical challenge in agriculture, causing substantial yield losses and threatening food security. In this work, we propose a hybrid deep feature engineering framework that integrates deep learning-based feature extraction with classical machine learning for accurate plant disease detection. A pretrained vision transformer (ViT) model is employed to extract discriminative features from leaf images, effectively capturing complex spatial relationships. To address the curse of dimensionality, principal component analysis (PCA) is applied, retaining 98% of the variance while reducing feature space complexity. The refined features are then classified using a support vector machine (SVM) optimized through hyperparameter tuning. Experimental results on the bean leaf lesions dataset demonstrate strong performance, achieving 92% accuracy and a weighted F1-score of 0.92. The proposed ViT–PCA–SVM pipeline effectively balances accuracy, computational efficiency, and generalization, making it a promising solution for real-time smart farming applications.

Why it matches plant phenotyping methods葉画像から植物病害状態を推定するViT–PCA–SVM解析パイプラインが研究の中心であり、植物表現型（病斑・病害状態）の画像ベース推定手法に該当する。

titleTransformer-based hybrid classification for plant leaf disease detection using vision transformer, principal component analysis, and support vector machine

High-throughput phenotyping of wheat ear surface area and ear density in the field →

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.

Contrastive multi-view representation learning for multi-camera plant phenotyping: A cotton field study →

Attempts to deploy computer vision in agricultural tasks often suffer from a shortage of annotated data. One strategy to alleviate the impact of limited data is Self-Supervised Learning (SSL), which involves pre-training a model on a pretext task that utilizes automatically generated annotations. The primary objective of this study is to leverage a multi-camera view dataset of cotton boll images for contrastive learning in order to enable phenotyping tasks with minimal data annotation. This dataset was collected in the field using six camera views. The efficacy of two contrastive learning frameworks (SimCLR and MoCo) in producing representations when positive examples originate from different cameras was investigated, and a comprehensive study of how the camera positions affect performance was conducted. After self-supervised pre-training, linear evaluation and semi-supervised learning experiments were performed on boll detection and plot status downstream tasks. In general, using multiple camera views with SimCLR and MoCo improves cotton boll detection mean average precision by 14% compared to vanilla SimCLR and MoCo. Through careful investigation using synthetic data, it was determined that relative camera poses with an intermediate amount of overlap seem more likely to perform well. Neither MoCo nor SimCLR was consistently superior to the other in this context. The representations embed meaningful features about the cotton plants, such as overall boll density, but also less meaningful ones, such as lighting variations. This technique could potentially accelerate the development of phenotyping algorithms based on data collected from field robots. • A contrastive learning method based on comparing multi-camera views was developed. • The method was tested with images of cotton bolls from a ground robot. • The method outperformed baseline contrastive learning approaches.

Why it matches plant phenotyping methodsマルチカメラ画像とコントラスト学習による植物表現学習・フェノタイピング手法の開発と評価が中心であり、綿花のボール検出性能を検証している。

abstractThe primary objective of this study is to leverage a multi-camera view dataset of cotton boll images for contrastive learning in order to enable phenotyping tasks with minimal data annotation.

Remote sensing data and machine learning models estimate sorghum grain yield in a plant breeding program →

P henotyping remains a critical bottleneck in sorghum ( Sorghum bicolor L. Moench) breeding programs, limiting rates of genetic gain due to labor-intensive yield estimation methods. To address this concern, this study investigates the potential of integrating remote sensing data with machine learning (ML) and deep learning (DL) models to improve sorghum grain yield predictions. Unmanned aircraft systems (UAS)-based imagery was collected across multiple field trials, extracting standard vegetation indices, canopy height features, and panicle traits using a YOLOv11-based object detection model, "YOLO-SORG." Six ML models-including ridge regression (RR), elastic net (EN), LASSO regression (LR), support vector regression (SVR), random forest (RF), and XGBoost (XGB)-were trained to predict plot-level yield using three distinct feature sets: panicle traits, canopy traits, and a combination of both. Results indicate that models relying solely or partially on canopy-derived features provided the most consistent and accurate yield estimates (R 2 ≈ 0.74-0.76), whereas models relying solely on panicle traits performed poorly (R 2 ≈ 0.28-0.42), indicating nadir-derived panicle metrics were potentially being indirectly captured with the canopy traits. Traditional regression models outperformed tree-based ensemble methods in variance partitioning and repeatability ( R ≈ 0.59-0.60), making them more suitable for many breeding applications. These findings highlight the promise of UAS-driven ML pipelines for non-destructive yield prediction but underscore potential limitations of nadir imagery for capturing panicle morphology and use in a robust yield prediction model. Future research should explore the inclusion of multi-temporal imaging, refined feature extraction approaches, and use of oblique, non-nadir imagery to enhance predictive accuracy in sorghum breeding programs.

Why it matches plant phenotyping methodsUAS画像からキャノピー高、穂形質、植生指数を抽出し、機械学習でソルガムのプロット収量を推定するパイプラインが研究の中心であり、形質取得・推定手法の評価も行っている。

abstractthis study investigates the potential of integrating remote sensing data with machine learning (ML) and deep learning (DL) models to improve sorghum grain yield predictions.

OPAL-Flow: Orientation-aware rice panicle detection and minute-scale anthesis rhythm identification under field conditions →

Accurate timing of rice panicle anthesis is critical for quantifying sterility risk under heat and humidity, yet minute-scale field measurement remains challenging because anthesis is transient and spikelets are tiny and difficult to detect. To address this, we present OPAL-Flow, a pipeline that provides single-panicle anthesis start and peak times under field conditions, consisting of a detector for panicle detection and tracking, a super-resolution reconstruction model, and an event-time pinpointing model. For slender panicle detection and panicle pose normalization, YOLO-SnakePanNet was introduced by using Dynamic Snake Convolution with a lightweight box-rotated head. Ablation experiments show that YOLO-SnakePanNet achieved mAP@50 of 94.4%, improving by 3.6% over the YOLOv11 while reducing computation by 0.7 GFLOPs. For panicle-level anthesis pinpointing, PanicleTimeMAE was proposed by incorporating a pyramid-dilated temporal convolutional network and a confidence-aware smoothing gate into the transformer, reaching Acc@±1 of 0.85 on 5-min sampled sequences (±1 frame = ±5 min), yielding a 40% decrease in MAE over VideoMAEv2. Finally, correlation analysis between variety-level anthesis start time (T start ) and peak time (T peak ) and same-day meteorology showed that higher photosynthetically active radiation (r = -0.543/-0.573 for T start /T peak ) and temperature (r = -0.288/-0.272) advanced anthesis, whereas higher relative humidity (r = 0.397/0.438) and rainfall (r = 0.428/0.502) delayed anthesis. The variance decomposition within fixed-effects model for Tstart ( R2 = 0.651) and Tpeak ( R2 = 0.648) prediction shows that variance mainly attributed to meteorological effects (64%) and variety effects (33.5%). Overall, OPAL-Flow enables variety selection for heat- and humidity-resilient anthesis in rice breeding and supports ecophysiological dissection of anthesis regulation.

Why it matches plant phenotyping methodsイネ穂の開花時刻という植物形質を圃場画像・動画から推定する検出、追跡、超解像、時刻推定パイプラインを開発し、性能評価も行っているため、植物フェノタイピング手法が研究の中心である。

abstractwe present OPAL-Flow, a pipeline that provides single-panicle anthesis start and peak times under field conditions, consisting of a detector for panicle detection and tracking, a super-resolution reconstruction model, and an event-time pinpointing model.

Ecological insights from transferable plant biomass mapping across the arctic using high-resolution structure-from-motion and LiDAR data →

Abstract Warmer temperatures, permafrost thaw, and increased wildfire activity are driving rapid ecological change across the Arctic, significantly altering plant productivity and aboveground biomass (AGB). These rapid changes highlight the urgent need to improve monitoring of vegetation dynamics in the Earth’s northern ecosystems, where high spatiotemporal heterogeneity occurs at scales finer than those captured by traditional satellite observations. The growing use of unoccupied aerial systems (UASs) presents an opportunity to overcome this limitation. Yet, the diversity of UAS platforms, sensors, and data collection and processing workflows presents challenges for developing standardized, generalizable approaches. To address this challenge, we compiled 672 AGB plots co-located with 183 UAS-based structure-from-motion (SfM) or light detection and ranging (LiDAR) surveys collected across the Arctic. Here, we: (1) evaluated the generalizability of UAS-derived canopy structure derived from high-resolution SfM and LiDAR for estimating AGB, (2) assessed scaling errors and their sources in two recent satellite-based AGB products derived from Landsat and moderate resolution imaging spectroradiometer, and (3) demonstrated the use of high-resolution AGB maps to quantify biomass variation across tundra plant functional types (PFTs) and to monitor post-fire recovery. Our results show that both SfM and LiDAR accurately captured AGB and its variability across tundra PFTs using a random forest model (overall root mean squared error: 0.332 kg m –2 ), with mapping performance varying slightly by region and data source. Using UAS-derived AGB maps as a benchmark, we identified systematic biases in satellite-derived AGB products, largely attributable to the magnitude of AGB and structural heterogeneity within coarse-resolution pixels. Applying our model to repeat UAS surveys following a tundra fire on Seward Peninsula, we observed rapid AGB recovery in non-shrub patches, with biomass recovering to pre-fire levels within two years. In contrast, shrub patches recovered more slowly, with AGB gains continuing over 2–4 years through both in-patch growth and lateral expansion (via dispersal) into remaining burned areas. Overall, these findings support the generalizability of UAS-based SfM and LiDAR data for estimating tundra AGB and highlight the need for broader collection and synthesis of such data to improve ecological monitoring and model benchmarking in the Arctic.

Why it matches plant phenotyping methodsUASのSfMおよびLiDARから植物群落の地上部 biomass (AGB) を推定する手法の一般化性能を評価し、衛星推定値のベンチマークにも用いており、植物形質取得が研究の中心である。

abstractevaluated the generalizability of UAS-derived canopy structure derived from high-resolution SfM and LiDAR for estimating AGB

TDD-YOLO: A novel model for precise detection of tomato diseases. →

Tomato diseases pose a significant threat to global agricultural production, often leading to substantial yield loss and major economic damage. Traditional disease detection methods rely on manual inspection, which is not only time-consuming and labor-intensive but also difficult to implement for real-time monitoring. While deep learning-based object detection techniques offer a potential alternative to manual inspection, existing models still face challenges in extracting subtle disease features, suppressing complex background interference, and in handling multi-scale disease representations in complex agricultural environments, limiting detection performance. To address these limitations, this paper proposes a novel TDD-YOLO model for precise tomato-disease detection (TDD) in complex agricultural settings. The proposed model is based on YOLOv11 with the following three main improvements: (1) a feature enhancement module is added to improve the backbone's ability to extract disease spot textures; (2) a joint attention mechanism is introduced to explicitly model cross-dimensional dependencies, effectively suppressing background interference; and (3) a feature fusion module is added to retain disease information across different scales while reducing computational costs. Experimental results, obtained on the Tomato-Village dataset (containing field-acquired images of tomato leaves with six diseases, collected in real agricultural environments, featuring complex backgrounds and varying illumination conditions) and Tomato-Disease dataset (emphasizing a greater diversity in tomato disease types along with healthy leaf samples), demonstrate that the proposed TDD-YOLO model outperforms the baseline in detection of tomato diseases (e.g., by improving mAP@50 and mAP@50:95, averaged across disease categories, by 4.1% and 6.0% on Tomato-Village and by 3.6% and 3.9% on Tomato-Disease, respectively) and state-of-the-art models (e.g., by improving the average mAP@50 and mAP@50:95, compared to the first runner-up, by 3.2% and 4.7% on Tomato-Village and by 2.4% and 2.1% on Tomato-Disease, respectively), while maintaining good parameter count and computational complexity, confirming its effectiveness and potential for practical usage in complex agricultural environments. The author-generated code and weight files are publicly available at https://github.com/LingShaQ/TDD-YOLOCode.

Why it matches plant phenotyping methodsトマト葉の病斑・病害状態を画像から検出するYOLOモデルを開発し、複数データセットでベースラインおよび既存モデルと比較検証しており、植物病害フェノタイピング手法が中心である。

abstractExperimental results, obtained on the Tomato-Village dataset

Quantifying the reliability gap in cross-domain plant disease classification: benchmarking the limited efficacy of standard mitigation techniques under controlled-to-field shift →

Introduction Confidence calibration, selective prediction, out-of-distribution scoring, and deep ensembles are mature techniques in machine learning, yet their efficacy under the severe domain shift encountered when plant disease classifiers move from controlled laboratory imagery to heterogeneous field photographs has not been systematically benchmarked. Methods Models trained on PlantVillage were evaluated on PlantDoc leaf-level crop images under a parent-image-aware split protocol, and a suite of standard mitigation techniques was applied to characterize the reliability gap. Analyses included temperature scaling and selective prediction for a fine-tuned ResNet-50, quantitative image-level shift analysis, Grad-CAM visualization, simple target-aware adaptation baselines, frozen-feature backbone comparisons, and ensemble baselines. Results In the primary case study, a fine-tuned ResNet-50 suffered a 67.7-percentage-point accuracy collapse upon cross-domain transfer, while mean predicted confidence remained at 79.76%. Post-hoc temperature scaling reduced calibrated ECE to 0.3645 but left selective risk at 80% coverage at 64.30%. Quantitative image-level shift analysis confirmed large-effect-size differences in saturation ( d = 3.90), border edge density ( d = 3.33), and foreground-occupancy proxy ( d = 2.48) between the two domains, while Grad-CAM visualizations showed that the model shifts attention from lesion-centered regions in PlantVillage to background-dominated areas in PlantDoc. Simple target-aware mitigations, including adaptive batch normalization and feature moment matching, improved accuracy from 0.321 to 0.343 and 0.366, respectively, whereas DANN-style adversarial adaptation degraded performance to 0.252. A frozen-feature backbone comparison across five backbones showed that, within the energy-scoring frozen-backbone comparison, DINOv2-S/14 achieved the highest unknown-detection AUROC (0.764) and the lowest selective risk at 80% coverage (0.520), with paired Wilcoxon tests confirming statistically significant accuracy and macro-F1 differences across backbones. Two ensemble baselines were evaluated: a warm-start end-to-end ResNet-50 ensemble reduced calibrated ECE to 0.063 but achieved only 0.666 AUROC, while a lightweight DINOv2 linear-probe ensemble achieved 0.779 AUROC after calibration but under limited epistemic diversity. Discussion Neither ensemble established deployment-grade reliability: the best selective risk at 80% coverage across all configurations remained above 0.51. The principal contribution is a reproducible, deployment-oriented reliability characterization showing that standard post-hoc and lightweight adaptation techniques reduce but do not eliminate the severe reliability gap under controlled-to-field transfer in agricultural computer vision.

Why it matches plant phenotyping methods植物病害画像分類の信頼性・ドメインシフト・校正・選択的予測を体系的にベンチマークしており、病害状態を画像から推定する方法の技術評価が中心である。

abstracttheir efficacy under the severe domain shift encountered when plant disease classifiers move from controlled laboratory imagery to heterogeneous field photographs has not been systematically benchmarked.

CNN MODEL OPTIMIZATION USING MULTI-STAGE DATA AUGMENTATION FOR LOCAL PLANT LEAF DISEASE CLASSIFICATION →

Plant leaf diseases are a major factor in reducing agricultural productivity, particularly for local commodities that often lack adequate artificial intelligence-based disease detection systems. This study aims to optimize the performance of a Convolutional Neural Network (CNN) model using the Inception V3 architecture through the application of multi-stage data augmentation to improve the classification accuracy of local plant leaf diseases. The dataset used is PlantifyDR from Kaggle, which has limited data volume and visual variation, requiring an effective augmentation strategy to improve the model's generalization ability. The proposed multi-stage augmentation approach consists of three stages—geometric, photometric, and texture-noise augmentation—that systematically enrich the diversity of training images. Evaluation results show that the proposed model provides significant performance improvements compared to the baseline model. The Inception V3 model with multi-stage augmentation achieved an accuracy of 0.762, an F1-score of 0.727, and a perfect AUC (1.00) across all classes, while the baseline model only achieved an accuracy of 0.595 and an average AUC of 0.877. Accuracy, loss, ROC curve, and confusion matrix analyses confirmed that multi-stage augmentation reduced overfitting and enhanced the model's ability to differentiate disease symptoms across leaf types. Therefore, this study concludes that multi-stage data augmentation is an effective approach for optimizing deep learning models on small and complex datasets, while also providing a significant contribution to the development of more accurate and reliable AI-based plant disease detection systems.

Why it matches plant phenotyping methods葉画像から植物病害症状を分類するCNNと、性能改善のための多段階データ拡張を開発・評価しており、植物の病害状態を推定する画像ベースのフェノタイピング手法が中心です。

abstractThis study aims to optimize the performance of a Convolutional Neural Network (CNN) model using the Inception V3 architecture through the application of multi-stage data augmentation to improve the classification accuracy of local plant leaf diseases.

Optimizing CNN-Based Transfer Learning through Fine-Tuning and Adaptive Augmentation for Chili Plant Disease Detection →

Chili peppers (Capsicum annuum L.) are a strategic horticultural commodity in Indonesia, but their productivity is often hampered by pathogen infections that cause leaf diseases such as anthracnose, leaf spot, and yellow virus. Early detection by farmers is still dominated by subjective visual observation and prone to misdiagnosis due to the similarity of symptoms between diseases. Although Deep Learning technology through Convolutional Neural Networks (CNN) offers an automated solution, implementation in real-world conditions still faces significant challenges such as lighting variations, complex backgrounds, and limited local datasets. This often leads to a drastic decrease in model performance compared to testing in a controlled environment. To address these issues, this study proposes an optimization of the transfer learning strategy on the MobileNetV2 architecture by integrating progressive layer-wise fine-tuning and adaptive data augmentation techniques. The fine-tuning method is carried out gradually on the pre-trained model layers, while adaptive augmentation dynamically manipulates images based on environmental characteristics to improve model robustness. The results of this study, which include multi-class classification on cross-location image data, are projected to be able to boost the accuracy and generalization ability of the model in heterogeneous field conditions. Practically, this research provides a framework for a more precise and robust disease detection system to accelerate the implementation of precision agriculture in the future.

Why it matches plant phenotyping methods植物葉の病徴を画像から分類するCNN手法の改良が研究の中心であり、病害状態の表現型推定に該当する。転移学習のファインチューニングと適応的画像拡張、異なる圃場条件での頑健性評価を扱っている。

abstractThe results of this study, which include multi-class classification on cross-location image data, are projected to be able to boost the accuracy and generalization ability of the model in heterogeneous field conditions.

3D Reconstruction and Knowledge Distillation to Improve Multi-View Image Models to Explore Spike Volume Estimation in Wheat →

Accurate estimation of wheat spike volume is important for yield component analysis and stress resilience assessment, yet field-based measurement remains challenging. Active 3D sensing methods such as Light Detection and Ranging (LiDAR) or time-of-flight (ToF) are sensitive to plant motion or poorly suited to outdoor conditions, while 3D reconstructions are computationally expensive. Direct 2D image processing would offer computational advantages, but image-based models lack explicit geometric information. We therefore propose a hybrid 2D-3D approach with knowledge distillation during training while enabling efficient image-only inference. First, we train a rigid-invariant point cloud network using distance-based histogram features to obtain pose-robust geometric representations. We then combine the 3D model with a proposed multi-view image-based regulated Transformer (RT) in an ensemble architecture. Finally, we distill the ensemble knowledge into a purely image-based student model using either feature-based or label-based distillation. The two distilled RTs reduce the mean absolute error (MAE) from 654.31 mm$^3$ of the non-distilled RT to 639.93 mm$^3$ and 644.62 mm$^3$, and increase correlation from 0.76 to 0.77 and 0.82, respectively. At the same time, inference time is reduced from 160 ms to 1.4 ms per spike. Distillation further mitigates volume-dependent bias and reshapes the latent representation of the image model toward a geometry-aware shape. Our results demonstrate that 3D-informed training of a 2D Transformer allows for scalable and efficient spike volume estimation for high-throughput field phenotyping.

Why it matches plant phenotyping methods小麦穂の体積を画像・3D再構成・知識蒸留で推定する手法の開発と性能評価が中心であり、高スループット植物フェノタイピングへの応用も明示されている。

abstractWe therefore propose a hybrid 2D-3D approach with knowledge distillation during training while enabling efficient image-only inference.

Integrating deep learning and field validation into a decision support system for Northern Corn Leaf Blight management in maize →

Northern Corn Leaf Blight (NCLB; also, Turcicum Leaf Blight, TLB), caused by Exserohilum turcicum (teleomorph: Setosphaeria turcica), is one of the most destructive foliar diseases of maize worldwide, often causing severe yield losses under favorable conditions. We developed a maize-specific, web-based Decision Support System (DSS) for real-time NCLB detection and management ( https://maize-nclb.streamlit.app/ ), integrating advanced deep-learning for automated diagnosis and fungicide advisory. Among thirteen Machine-learning and deep-learning models evaluated for classification, the Visual Geometry Group 16-layer convolutional neural network (VGG16) outperformed all others, achieving 94.0% accuracy, with balanced precision, recall, and F1-score of 0.94, and an AUC-ROC of 0.93. Confusion matrix analysis revealed minimal misclassification, with only 12 errors out of 357 samples, confirming the model's high reliability in distinguishing healthy and infected plants, while Grad-CAM visualizations consistently highlighted biologically meaningful lesion regions, supporting the model's interpretability and alignment with plant pathological symptoms. Field validation of DSS-guided fungicide recommendations (Azoxystrobin 18.2% + Difenoconazole 11.4% w/w SC) demonstrated significant benefits, reducing disease incidence to 6.8% compared with 67.4% in controls, achieving 90% disease reduction, and enhancing grain yield by 35.4% (8.55 t/ha), with a favorable cost-benefit ratio of 1:2.49. Seasonal disease progression analysis further confirmed DSS effectiveness, with cumulative disease burden reduced by approximately 85% compared with untreated control. These results highlight the potential of integrating deep-learning with field-validated management strategies into a practical DSS, demonstrating its potential for precision disease management in maize.

Why it matches plant phenotyping methods葉の病斑を画像から分類・可視化する深層学習法を開発し、野外で検証した研究であり、植物病害状態のフェノタイピング手法が中心です。

abstractintegrating advanced deep-learning for automated diagnosis and fungicide advisory

LeafyVGG-16: Transfer Learning for Plant Disease Detection with Cyber Risk Analysis →

Plant disease detection using deep learning is essential for precision agriculture, enabling early and automated crop health monitoring. This study proposes an end-to-end transfer learning pipeline, LeafyVGG-16, for multi-class classification of plant diseases and nutrient deficiencies using a tomato leaf dataset. The framework integrates data preprocessing, augmentation, and a VGG-16 backbone with a two-stage fine-tuning strategy. The proposed model is evaluated against CNN, DenseNet-121, Inception-V3, EfficientNetB0, and ResNet-50, achieving an accuracy of 0.93 with precision, recall, and F1-scores of 0.93, 0.90, and 0.92, respectively. These results demonstrate the effectiveness of transfer learning for fine-grained plant disease recognition. We further evaluate model robustness under adversarial cyber attacks to assess deployment reliability in agricultural systems. Under Fast Gradient Sign Method (FGSM) attacks ( ϵ = 0.01– 0.05), the model shows an accuracy drop of 1%–7.5%, while Projected Gradient Descent (PGD) attacks ( ϵ = 0.05, step size = 0.005, 10 iterations) produce similar degradation, highlighting the model’s vulnerability to adversarial perturbations. These findings highlight potential security and reliability risks in AI-based agricultural decision-making systems. Future work will focus on improving robustness and cyber-resilience and extending this framework to other crops for secure and context-aware deployment in resource-constrained environments.

Why it matches plant phenotyping methods植物葉の画像から病害・栄養欠乏状態を分類する深層学習パイプラインが研究の中心であり、複数モデルとの比較評価と敵対的攻撃下での頑健性検証も実施しているため。

abstractThis study proposes an end-to-end transfer learning pipeline, LeafyVGG-16, for multi-class classification of plant diseases and nutrient deficiencies using a tomato leaf dataset.

A multi-stage, pixel-level annotated apple dataset for precision agriculture research. →

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).

Adaptive fuzzy deep learning with multimodal sensor fusion for enhanced plant disease detection →

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.

ConvGeM-next: a deep learning framework for plant disease detection →

Introduction Plant diseases pose a major challenge to sustainable agriculture, particularly in regions that heavily depend on farming. Early and accurate identification of plant diseases is crucial for ensuring food production and minimizing crop losses. The rapid advancement of deep learning, particularly in convolutional neural networks (CNNs), has significantly enhanced plant disease classification performance. However, many models often struggle to generalize effectively in real-world scenarios due to challenges such as low-intensity visuals, low contrast between the background and foreground of the suspected sample, noise, and chrominance variation. Methods To address the challenges mentioned above, we introduce ConvGem-NeXt, an end-to-end deep learning architecture specifically designed for fine-grained plant disease classification, built on the ConvNeXt baseline model featuring enhanced generalization capabilities. More precisely, our method incorporates a learnable Generalized Mean pooling layer and ReLU activation in the ConvNeXt model to enhance spatial feature representation, and a custom classifier head that integrates batch normalization, ReLU activation, and dropout to mitigate overfitting and improve classification accuracy. Results We tested the presented model on two large-scale and diverse databases, PlantVillage and the PlantDoc. The model achieved 99.65% accuracy on the PlantVillage dataset and 94.69% accuracy on the real-world PlantDoc dataset, demonstrating the efficacy of our method for reliably classifying plant diseases. Discussion This work contributes to the rapidly growing field of agricultural automation by providing a reliable framework for timely disease diagnosis and supporting the enhancement of crop productivity.

Why it matches plant phenotyping methods植物病害を画像から分類する深層学習アーキテクチャを開発し、複数データセットで性能検証しており、植物の病害状態の取得・推定が研究の中心である。

abstractwe introduce ConvGem-NeXt, an end-to-end deep learning architecture specifically designed for fine-grained plant disease classification

GBR-DETR: A Real-Time Tomato Leaf Disease Detection Model for Edge Device Deployment. →

Tomato leaf diseases pose significant threats to crop yield and food security. However, in real-world cultivation environments, factors such as fluctuating illumination, varying leaf occlusion, and ambiguous lesion morphology often compromise detection accuracy. This paper presents the Gradient-aware Bidirectional Retentive Detection Transformer (GBR-DETR), a model designed for high-precision, real-time disease detection. This model is composed of two network structures and a retentive feature aggregation module: (1) a Multi-scale Gradient-Aware Transfer Network (MGAT-Net) is designed to encode gradient information through the Sobel operator, thereby enhancing the localization stability for small and blurry lesions; (2) a Bidirectional Context Pyramid Network (BCPN) is proposed to enable bidirectional interactions among multi-level features through a top-down and a bottom-up pathway, thereby generating multi-scale lesion features and bridging cross-scale semantic gaps; and (3) a Retentive Feature Aggregation Module (RFAM) is used to suppress background noise and establish global feature correlations, thereby enhancing the overall representation capability for lesion recognition. Experiments on the Multi-scenario Tomato Leaf Disease (M-TLD) dataset show that GBR-DETR yields gains of 3.12, 4.88, and 3.41 percentage points in mAP 50-95 , mAP 50 , and mAP 75 , respectively, over the baseline RT-DETR, while also outperforming representative DETR-based and CNN-based detectors. The model demonstrates robust generalization on the PlantDoc cross-domain benchmark, achieving a 2.11% improvement in mAP 50 over the baseline. Deployed on the NVIDIA Jetson Orin Nano with TensorRT FP16, it achieves 54 ms latency, enabling real-time disease monitoring on edge devices. This solution provides effective technical support for real-time disease monitoring in smart agriculture.

Why it matches plant phenotyping methodsトマト葉の病斑・病害状態を画像から検出するモデルを開発し、複数データセットで比較検証、エッジデバイス実装まで評価しており、植物病害表現型の取得手法が中心である。

abstractThis paper presents the Gradient-aware Bidirectional Retentive Detection Transformer (GBR-DETR), a model designed for high-precision, real-time disease detection.

CKM-YOLO11: A Lightweight Maize Foliar Disease Detection Model for Complex Natural Field Environments. →

Accurate and real-time detection of maize foliar diseases is important for field disease monitoring and yield protection. However, in complex natural field environments, different diseases often exhibit high visual similarity, and early weak lesions are easily confused with background elements such as dry leaves, soil, and shadows, leading to false positives and missed detections in existing models. To address these challenges, this study proposes an improved lightweight maize foliar disease detection model based on YOLO11, termed CKM-YOLO11. First, a mixed local channel attention mechanism is introduced and adapted to the task in the backbone to construct the C3k2-MLCA module, thereby enhancing joint modeling of local lesion textures, edge details, and global contextual information. Second, a lightweight residual attention module, named MLCA-HeadLite, is designed at the P5 layer of the neck/head to alleviate the suppression of weak lesion responses during deep feature fusion. Experimental results demonstrate that the proposed model achieves an mAP@50 of 81.5% on a self-constructed maize disease dataset with complex field backgrounds, improving mAP@50 and mAP@50-95 by 3.2 and 3.4 percentage points, respectively, compared with the baseline YOLO11, while maintaining a low parameter count and computational cost. Further analyses based on the confusion matrix, comparisons of detection results, and Grad-CAM visualizations indicate that the proposed model performs better in background suppression, retention of weak lesion responses, and robustness in complex scenes. This study provides a reference for the lightweight design of maize foliar disease detection models in complex field environments and their deployment on agricultural edge devices.

Why it matches plant phenotyping methodsトウモロコシ葉の病斑・病害状態を画像から検出する軽量モデルを開発し、データセット上で性能評価しているため、植物病害表現型の取得手法が中心である。

abstractthis study proposes an improved lightweight maize foliar disease detection model based on YOLO11, termed CKM-YOLO11.

The Significance of Hybrid CNN and ANN Model in Design and Implementation of Deep Learning Model for Plant Disease Detection →

Plant disease is a serious threat to agricultural productivity and food security worldwide. Traditional diagnostic methods such as manual observation and laboratory testing are time-consuming, labor-intensive and error prone. The emergence of artificial intelligence (AI) and deep learning (DL) offer scalable solutions for precision agriculture in plant disease detection using advanced computational techniques to process large datasets. Hybrid deep learning architecture integrates Convolutional Neural Networks (CNNs) along with Artificial Neural Networks (ANNs) can leverage both visual and contextual data to improve detection performance. The hybrid CNN-ANN model was developed to analyze visual data (plant images) and contextual data (environmental and soil metrics). The CNN module extracted spatial and textural features from plant images, while the ANN module processed environmental parameters. These outputs were fused into a unified feature vector for disease classification. A total of 15 plant species and their associated diseases were analyzed using 200-270 training samples and 150-190 testing samples for each disease across a total of 1000 images. The model was judged by metrics such as detection accuracy, AUC, sensitivity etc. Data augmentation, pre-trained architectures (e.g., ResNet50) and early stopping techniques were utilized to improvise model performance. The hybrid model saliently achieved detection accuracy consistently above 87% with majority of diseases surpassing 90%. Highperforming cases like Rice Blast (92.5%), Tomato Early Blight (93.8%), and Coffee Rust (93.0%), with AUC values of 0.93 or higher, sensitivity exceeding 94% and specifically above 90%. Diseases of Sugarcane Red Rot and Tea Blister Blight exhibited sensitivities of 92.4% and 92.1% and specificities of 91.1% and 90.5% respectively. Moderate accuracy for Coconut Bud Rot (87.5%) and Mustard Alternaria Blight (87.8%) was due to smaller training sample sizes.

Why it matches plant phenotyping methods植物画像から病害状態を分類するハイブリッドCNN-ANN手法を開発し、精度・AUC・感度などで評価しており、病害フェノタイピング手法が研究の中心である。

abstractThe hybrid CNN-ANN model was developed to analyze visual data (plant images) and contextual data (environmental and soil metrics).

Research on Apple Surface Disease Detection Method Based on Improved YOLOv11s. →

Apple surface diseases are crucial factors affecting the quality and yield of apples. Traditional manual inspection methods suffer from low efficiency and poor real-time performance. To address these issues, this paper proposes an apple surface disease detection method based on an improved YOLOv11s. Firstly, three groups of GAM attention mechanisms are integrated into the neck structure of the YOLOv11s to enhance the efficiency of feature fusion and the capability of semantic information transmission. Secondly, the original convolutional downsampling in the backbone network is replaced with a Haar-based feature downsampling module, enabling the model to retain more high-frequency detail information during the downsampling process. In addition, the WFU module is introduced to realize the dynamic allocation of feature weights, enhancing the model's ability to recognize multi-scale defect features. Finally, the PIOUv2 loss function is adopted to optimize bounding box regression, improving the model's detection performance for tiny defect spots. In addition, various data augmentation methods for small datasets are employed to improve the model training performance and effectively avoid the problem of data overfitting. The experimental results demonstrate that the F1-score of the proposed model is increased by 4.2%, and the mAP@50:95 is boosted by 2.4%. The detection performance outperforms various comparative models, which verifies the effectiveness and superiority of the proposed method.

Why it matches plant phenotyping methodsリンゴ表面の病斑・欠陥を画像から検出する改良YOLO手法の開発と比較検証が研究の中心であり、植物の病害状態を直接推定するフェノタイピング手法に該当する。

abstractthis paper proposes an apple surface disease detection method based on an improved YOLOv11s.

AgroDualNet: a dual deep learning-based crop disease forecasting and fruit ripening detection. →

Proper diagnosis of crop diseases and accurate measurement of fruit ripeness is essential in enhancing agricultural productivity, but conventional methods of diagnosis are time-consuming, error-prone, and inefficient. With the rapid development of AI, deep learning (DL), and IoT, there is increasing demand for combined solutions that jointly address plant health monitoring and harvest optimization in a reproducible and deployment-oriented manner. This study develops a new bi-phasic DL framework, AgroDualNet, that predicts crop diseases and identifies fruit ripeness stages to optimize yield quality and minimize agricultural losses. The work explicitly targets improved classification reliability, broader class evaluation, rigorous validation and generation of decision-ready outputs for precision agriculture. AgroDualNet comprises two modules. The crop-disease prediction module integrates ResNet50 with a Convolutional Block Attention Module (CBAM), and a Sequential Minimal Optimization (SMO)-based SVM classifier to enhance feature learning and classification performance Several different architectural designs are benchmarked and the resultant model is tested on both a dedicated 3-class subset and a large multi-class model of the PlantVillage dataset with leakage safe protocol(augmentation applied only on training data), cross-validation, statistical significance testing as well as ablation. The fruit-ripeness module employs YOLOv8 for real-time fruit localization and MobileNetV2 for lightweight ripeness classification suitable for edge deployment and a prototype decision-support layer maps predictions to actionable recommendations. That is able to run on the edge. Experiments show that the hybrid CBAM + ResNet50 + SMO model achieves 99.6% accuracy for crop disease classification on a three-class configuration of the PlantVillage dataset and maintains consistently higher accuracy than strong baseline in a 38-class setting, with statistically significant results confirmed by McNemar's test (p < 0.001) outperforming baseline and intermediate architectures in accuracy, precision, Recall and F1-Score The fruit ripeness pipeline achieves 98.88% classification accuracy across four ripeness stages (unripe, semi-ripe, ripe, over-ripe) on a combined Kaggle and real-field apple dataset with low inference time, confirming its suitability for near real-time deployment on edge devices. Cross-validation, Statistical significance tests and ablation studies collectively validate the robustness and significance of these gains and the decision-support layer demonstrates the feasibility of converting raw predictions into interpretable, recommendation-oriented outputs. AgroDualNet provides an efficient and unified system for monitoring plant diseases and evaluating fruit ripeness, with statically validated performance across both focused and full multi-class settings, addressing two critical challenges in precision agriculture with a single extensible framework. The dual-module design of AgroDualNet, which combines disease prediction with ripeness analysis and a preliminary decision-support prototype offers a more comprehensive and practically relevant AI-driven monitoring solution than conventional single-task models. By emphasizing multi-class validation on PlantVillage, leakage-aware experimentation, statistical verification, and system-level integration, this works supports real-time, precise and automated guidance to reduce crop losses, improve harvest timing, and enable smarter farm-level decision making.

Why it matches plant phenotyping methods植物病害状態と果実成熟度を画像から推定する深層学習パイプラインの開発・比較検証が中心であり、PlantVillageおよび実圃場データで交差検証、アブレーション、統計検定を実施しているため、植物フェノタイピング手法として含める。

abstractThis study develops a new bi-phasic DL framework, AgroDualNet, that predicts crop diseases and identifies fruit ripeness stages

A Low-cost "Plant-Scanner" Platform for Automated Detection of Ustilago Maydis Infection in Maize Using Deep Learning →

Abstract Ustilago maydis is a biotrophic fungus that causes smut disease in maize, leading to tumor formation on aerial parts of the plant. While U. maydis has been a model for plant-fungal interaction studies, no tool has existed to automatically quantify infection symptoms under laboratory conditions for deep learning analysis. To address this, we developed a rotating camera system that captures videos of plants under customized lighting and shutter settings. These videos were used to train machine learning models to distinguish between healthy and infected plants. Two machine learning models have been presented. In the first approach, by employing a naive masking technique and combining classical machine learning with deep learning classifiers, the model achieved a reasonable performance, with an Area Under the Curve (AUC) of 0.90 on the Receiver Operating Characteristic (ROC), displaying high sensitivity and specificity. The second approach utilizes pre-trained YOLO11 model for object detection and further classification. The YOLO11-based approach outperforms traditional methods, achieving near-perfect accuracy (AUC: 0.99-1.00), demonstrating its superiority for real-time, scalable applications. Our toolset, featuring a cost-efficient and customizable scanning platform with open building-blocks design, provides a valuable resource for unbiased disease symptom detection and scoring, with potential applications in other plant pathology studies. This point enables easy replication and adaptation by other research laboratories which makes the platform robust, scalable and practical beyond our specific application.

Why it matches plant phenotyping methods植物の感染症状を画像から自動検出・定量する低コスト撮像プラットフォームと解析モデルを開発しており、植物表現型取得法が中心的である。

abstractwe developed a rotating camera system that captures videos of plants under customized lighting and shutter settings.

Optimized convolutional neural network using bacterial colony optimization for plant leaf disease detection →

The currently growing effects that plant diseases have on global agriculture require the creation of highly intelligent and accurate detection systems. Convolutional Neural Networks (CNNs), as deep learning, have proved effective in the detection of plant diseases based on images. The CNN performance, however, is very sensitive to hyperparameter tuning, which usually requires manual, sub-optimal tuning. The study suggests a new method to detect plant diseases with an optimized CNN architecture optimized by the Bacterial Colony Optimization (BCO). The BCO algorithm replicates the adaptive foraging behavior of bacterial colonies to automatically determine optimal CNN hyperparameters, such as the number of filters, kernel size, pooling methods, learning rate, and dropout probability of the CNN. The experiment conducted on the PlantVillage dataset demonstrated that the proposed BCO-CNN achieved an accuracy of 96.68% with a false alarm rate (FAR) of 4.05% outperforming other heuristic-fitted models such as particle swarm optimization (PSO) - CNN, CNN with support vector machine (SVM), and CNN, VGG16 in accuracy, precision, recall, and F1-score. The given work offers an automated, scalable approach to the accurate, early detection of the diseases of the plant, contributing to better yields of crops and sustainable agriculture.

Why it matches plant phenotyping methods植物葉画像から病害状態を推定するCNNのハイパーパラメータ最適化手法を開発・評価しており、植物フェノタイピング手法が研究の中心である。

abstractThe study suggests a new method to detect plant diseases with an optimized CNN architecture optimized by the Bacterial Colony Optimization (BCO).

Autonomous Embedded-Vision System for Multistage Detection of Phytopathogenic Fungi in Potato and Tomato Crops UsingConvolutional Neural Networks →

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

Deep learning techniques for early detection and classification of leaf diseases in crops. →

Introduction Rapid population growth and climate change have intensified the need for sustainable agricultural productivity. Plant leaf diseases significantly impact the crop yield, quality, and food safety, necessitating accurate and automated detection methods. Methods This study proposes a deep learning (DL)-based framework for automated detection and classification of tomato and soybean leaf diseases. The proposed framework is trained and evaluated over a large-scale datasets comprising 16,012 tomato leaf images and 6,410 soybean leaf images. Multiple convolutional neural network (CNN) models, including DenseNet121, MobileNetV2, and InceptionV3, are employed for classification. Object detection is performed using YOLOv12. To enhance interpretability, Gradient-Weighted Class Activation Mapping (Grad-CAM) is integrated. Furthermore, a novel Hybrid Attention-Based Stacking Ensemble Model is developed using ResNet152V2, VGG19, and EfficientNetB0, combined with Convolution Block Attention Module (CBAM) and spatial attention mechanisms. Results The CNN models achieved classification accuracies of 97% for DenseNet121, 98% for MobileNetV2, and 99.94% for InceptionV3. YOLOv12 attained a mean average precision (mAP) of 99.5%. The proposed hybrid ensemble model achieved an accuracy of 99.18%, demonstrating improved feature learning through combined channel and spatial attention. Grad-CAM visualizations confirmed that the model effectively identifies the disease-relevant regions. Discussion The results indicate that the proposed framework has attained a high accuracy, robustness, and interpretability for plant disease detection. The integration of attention mechanisms and explainable AI enhances model reliability and transparency. This framework shows a strong potential for the real-time agricultural monitoring, although further validation across diverse crops and real-world field conditions is required.

Why it matches plant phenotyping methods植物葉の病害状態を画像から自動検出・分類する深層学習フレームワークの開発であり、病害表現型の取得・推定が研究の中心。

abstractThis study proposes a deep learning (DL)-based framework for automated detection and classification of tomato and soybean leaf diseases.

Advanced deep learning vision transformer models for intelligent grain counting in agricultural data analytics. →

Grain number estimation plays a crucial role in agriculture, serving as a key indicator for crop yield and quality assessment. With advances in computer vision, automatic grain detection has become a significant research area, where deep learning methods have shown remarkable promise. This study proposes a vision transformer model called Swin Transformer, which leverages hierarchical attention mechanisms across shifted windows to effectively capture both local and global features of grains in complex imagery. The model achieves the highest accuracy of 98%, outperforming baseline traditional CNN (ResNet-50) and DINO models in grain counting tasks. To support and validate model performance, explainable AI (XAI) techniques such as Grad-CAM and LIME are employed, highlighting the interpretability and focus of the model on relevant grain regions. Furthermore, a comprehensive empirical analysis is conducted using multiple statistical tests to evaluate the model's robustness and generalizability across various grain morphological parameters, establishing the Swin Transformer as a powerful and interpretable solution for intelligent grain counting in agricultural data analytics.

Why it matches plant phenotyping methods画像から穀粒数を推定する深層学習手法の開発・比較検証が研究の中心であり、植物の収量関連形質を測定するため、植物フェノタイピング手法として収録する。

abstractThis study proposes a vision transformer model called Swin Transformer

BloomSight: An ultra-high-frequency phenotyping framework for diurnal flowering dynamics in japonica and indica rice to enable genetic dissection and hybrid-breeding applications. →

Rice ( Oryza sativa ) production underpins food security in many rice-consuming nations. As a critical developmental transition that directly determines yield and grain quality, flowering dates and timing are genetically complex and highly sensitive to environmental fluctuations. This complexity requires new methods to quantify diurnal floral characteristics, which are essential to hybrid breeding in cereals. Here, we present BloomSight, an ultra-high-frequency and deep-learning (DL) powered framework for phenotyping and measuring minute-level flowering dynamics in japonica and indica rice. After monitoring 172 rice accessions selected from the Chinese Rice Mini-Core Collection using cost-effective time-lapse imaging platforms for 16 days, we acquired over 530,000 accession-level images and established the Open Rice Flowering Training (ORFT) dataset, with over 39,000 panicles and 350,000 anthers annotated. Next, a two-stage customised DL model (i.e. YOLACT-Panicle for panicle segmentation and UNet-Anther for anther identification) was trained using the ORFT set, enabling ultra-high-frequency measures of anther extrusion at the minute level. Based on trait analysis, we further fitted curves to dynamically identify diurnal flowering patterns, including key timepoints such as the initial flowering timepoint ( T Ini. ), quickest flowering timepoint ( T Qck. ), and peak flowering time ( T Peak ), and novel traits such as the duration of rapid flowering phase ( P Rpd. ) and flowering density across key phases. After validating BloomSight-derived traits against manual observations, we classified the japonica and indica accessions into three patterns: Slow, Moderate, and Fast, all of which had distinct flowering windows. These analyses helped us integrate phenotypic variations into a genome-wide association study (GWAS), revealing many significant single nucleotide polymorphisms (SNPs) associated with known (e.g. EMF1 , OsMYB8 , and PME42 ) and several repeatedly identified unknown loci (one of these loci has been recently verified by other groups), demonstrating the value of the BloomSight framework. Taken together, we believe that BloomSight provides an ultra-high-frequency framework for diurnal flowering phenotyping, enabling the measurement of biological meaningful floral traits with minute-level resolution that can enable flowering-related developmental studies and hybrid-breeding applications in rice and more broadly benefit the plant and crop research community.

Why it matches plant phenotyping methodsイネの開花動態を高頻度画像と深層学習で抽出するフェノタイピング基盤を開発し、データセット構築と手動観測による検証も行っているため、方法が研究の中心である。

abstractwe present BloomSight, an ultra-high-frequency and deep-learning (DL) powered framework for phenotyping and measuring minute-level flowering dynamics in japonica and indica rice

Random Rolling Attention Augmentation for Efficient Agricultural Disease and Pest Detection →

Abstract Accurate detection of agricultural diseases and pests is essential for crop protection and food security worldwide. Real-world field applications face challenges including small objects, complex backgrounds, high class similarity, and limited computational resources. This work proposes a Random Rolling Transformer (RRT), which introduces random circular shifts along channel and sequence dimensions into multi-head self-attention to enrich feature interactions without increasing parameters or computation. Integrated into YOLOv12, the proposed RRT-YOLO is evaluated on the IP102 pest dataset and a tomato leaf disease dataset. Results show that RRT-YOLO improves mAP@50 by 2.5% and mAP@50–95 by 3.9% on IP102, and by 8.8% and 4.2% on the tomato disease dataset, while maintaining identical model size and complexity. This attention perturbation strategy offers an effective and efficient solution for lightweight agricultural vision detection and can be extended to other visual computing tasks. The code and detailed descriptions can be accessed via the following repository: \href{https://github.com/glorioustory/Random-Rolling-Attention-Augmentation.git}{https://github.com/glorioustory/Random-Rolling-Attention-Augmentation.git}.

Why it matches plant phenotyping methods植物葉の病害状態を画像から検出する計算手法を開発・評価しており、植物病害の表現型取得が技術的中心である。害虫検出も含むが、トマト葉病害データセットで手法性能を検証している。

abstractThis work proposes a Random Rolling Transformer (RRT), which introduces random circular shifts along channel and sequence dimensions into multi-head self-attention to enrich feature interactions without increasing parameters or computation.

Investigating performance and key factors for real-world deployment of grain image classification using convolutional neural networks →

Accurate and efficient grain quality assessment is critical for making informed decisions throughout the grain value chain. Early detection of disease enables actions to mitigate spread and further damage, and optimal batch mixing to fulfill specified quality requirements allows for maximizing value and minimizing scrapping. Vision based machine learning and deep learning approaches are gaining attention in the agricultural sector and are useful for the development of automated grain quality assessment. These techniques can reduce the current manual inspection load and are key for objective and precise analysis. Yet, the majority of prior studies are constrained to small or controlled and curated datasets. Practical challenges associated with real-world deployment and reliability are rarely addressed. That is the focus of this work. We present and demonstrate a structured approach for investigating convolutional neural networks (CNNs) and key factors influencing performance for wheat kernel classification. The objective is to determine a CNN model that ensures high and robust classification accuracy, while elucidating and explaining how different image dataset characteristics and training parameters affect performance and reliability. We use a commercial mirror-based imaging system that captures over 90% of each kernel's surface and contrast and compare model architectures, robustness, the effect on pre-processing and image resolution. Our results show similar and high overall performance for ResNet50V2 and EfficientNetV2B0 ([Formula: see text]% accuracy), but per-class analysis indicate that the smaller classes suffer from lack of representative examples, and that most classes benefit from pre-processing including downsampling whereas others benefit from higher resolution. Interactive visualizations reveal that another contributing factor is dubious annotation and multi-class belongingness. Thus, our step-by-step analysis of CNN performance underscores the need for representative data, proper pre-processing, and class-aware evaluation to ensure trustworthy deployment in wheat grain quality assessment.

Why it matches plant phenotyping methods小麦粒画像から品質・病害クラスを推定するCNN画像解析手法の性能、頑健性、前処理、解像度、データ特性を体系的に評価しており、フェノタイピング手法が中心的である。

abstractWe present and demonstrate a structured approach for investigating convolutional neural networks (CNNs) and key factors influencing performance for wheat kernel classification.

AF-RT-DETR: Adaptive cross-scale feature interaction for real-time plant disease detection in complex field environments →

Introduction Accurate plant disease identification is of great importance for ensuring agricultural productivity and food security. However, complex illumination variations, leaf occlusion, and diverse disease spot scales throughout plant growth stages significantly increase the difficulty of real-time detection, leading to limited accuracy and robustness in existing approaches. Methods To address these challenges, we propose an improved RT-DETRv2-based plant disease detection model, termed AF-RT-DETR. A Bidirectional Cross Gate (BCG) module is introduced in the feature extraction stage to reduce channel redundancy and enhance discriminative feature representation through multi-level feature interactions. The original RepVGG structure is replaced with a Dynamic Channel Shift (DCS) module, effectively enlarging the receptive field and strengthening contextual feature fusion without additional computational overhead. Additionally, an improved Scale-aware Multi-level Loss (SML) emphasizes low-quality feature maps to improve detector robustness. Results The model achieves mAP50 and mAP50:95 of 93.6% and 67.2% on the Plant-Disease dataset, surpassing the baseline by 5.1% and 4.5%. Furthermore, the model was evaluated on multiple crops and growth stages under diverse field conditions, demonstrating robust performance and adaptability. Discussion These results indicate that AF-RT-DETR effectively enables real-time plant disease detection in complex field environments.

Why it matches plant phenotyping methods植物の病徴を画像から検出するモデルの開発と、複数作物・生育段階・圃場条件での性能評価が中心であり、植物病害状態の表現型計測手法に該当する。

abstractwe propose an improved RT-DETRv2-based plant disease detection model, termed AF-RT-DETR.

TriAttnNet Based Deep Learning Model for Automated Cotton Pest Detection and Disease Classification →

Abstract This paper presents a deep learning model to detect cotton plant pests and classify diseases, which must overcome limited datasets, class imbalance, and feature redundancy. At the preprocessing phase, the Gaussian blur filtering and Contrast Limited Adaptive Histogram Equalization (CLAHE) are used to sharpen images by improving their clarity and contrast. To increase and diversify the data, Spa-GAN-based data augmentation is used to produce realistic synthetic samples. To obtain an accurate Region of Interest (RoI), an Attention-Guided Multi-Scale Residual U-Net (AGMS-UNet) is considered to segment local and global structural information. The proposed framework makes three major contributions: (i) a new attention-based feature extractor, TriAttnNet , that incorporates spatial, channel, and contextual attention to represent diseases on a fine-grained level; (ii) a new optimization strategy, Hybrid Mongoose Ray Chaotic Optimization (HMRCO), which includes chaotic strategies to better tune the parameters and explore the feature space; and (iii) classification layer with focal loss for final decision. Experimental analyses prove that the suggested method is much more effective than the current state-of-the-art models, providing a powerful and understandable solution to precision agriculture and sustainable cotton crop health monitoring. Experimental results show that TriAttnNet achieves 98.66% accuracy, 98.71% recall, and 98.81% F1-score, which is better than the state-of-the-art algorithms, such as EfficientNetB1-CBAM (96.38%) and BERT-ResNet-PSO (95.69%). The proposed system is computationally feasible and interpretable, and it is interpretable to provide a practical solution to precision agriculture and sustainable monitoring of the health of cotton crops.

Why it matches plant phenotyping methods綿植物の病害を画像から検出・分類する深層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して性能比較・検証しているため。

abstractThis paper presents a deep learning model to detect cotton plant pests and classify diseases

ODANet: an occlusion and density aware network for small object detection of coffee cherry ripeness in complex field environments. →

Introduction Coffee cherry ripeness assessment is critical for harvesting efficiency and product quality, yet traditional manual inspection methods suffer from subjectivity and low efficiency. Methods To address the challenges of detecting small, occluded, and densely distributed coffee cherries in complex field environments, this study proposes an Occlusion and Density Aware Network (ODANet). Built upon the YOLOv8 framework, ODANet integrates three innovative modules: (1) Condition-Guided Windowed Attention (CGWA), which incorporates occlusion and density maps as auxiliary guidance signals for efficient feature enhancement; (2) Attention-guided Space-Preserving Convolution (ASPC), which employs space-to-depth transformation with cascaded attention to preserve spatial information during downsampling; and (3) Dual-Adaptive Dynamic Upsampling (DADU), which achieves content-adaptive feature reconstruction through dual-branch offset prediction with learnable fusion weights. Results Comprehensive evaluation on a publicly available dataset demonstrates that ODANet achieves state-of-the-art performance among 17 diverse detection architectures, attaining 76.7% mAP@0.5 with a 6.3 percentage point improvement over baseline YOLOv8, while maintaining computational efficiency (8.1 GFLOPs, 30.4M parameters) suitable for real-time deployment. Ablation studies validate the contributions of each module: ASPC improves performance by 2.2%, DADU by 0.6%, and CGWA by 3.5%. Discussion The model demonstrates robust performance across varying lighting conditions, occlusion levels, and growth stages, making it particularly suitable for practical agricultural deployment. This research provides an efficient solution for small object detection in precision agriculture.

Why it matches plant phenotyping methodsコーヒーチェリーの成熟度という植物器官の状態を画像から推定する検出手法を開発し、複数モデル比較・アブレーションで技術的に検証しているため、植物フェノタイピング手法が中心である。

abstractthis study proposes an Occlusion and Density Aware Network (ODANet)

Convolutional Neural Networks for Detecting White Grape Clusters in High-Density Vineyards →

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.

Intelligent identification of rice leaf diseases via improved faster-RCNN with multi-feature scale fusion. →

Many Artificial Intelligence and Machine Learning technologies have been applied to detect rice diseases. These approaches are either unable to identify the diseases or have a slow recognition speed. Therefore, an improved Faster-RCNN (Faster-RCNN-Pro) model is proposed to overcome these issues. First, SENet attention modules are embedded in the backbone of Faster-RCNN to enhance confidence of objects that are difficult to recognize by enhancing key image information and suppressing background information. Second, structure of the feature extraction network and RPN are improved by using multi-feature scale fusion to increase the utilization of micro-target features. Third, the quantization error introduced in the process of pooling the region of interest is then eliminated by ROI Align. Finally, a balanced L1 loss function is designed to effectively reduce the imbalance between samples with a large gradient that are difficult to learn, and samples with a small gradient that are easy to learn. The experiment results show that the improved model has a better detection accuracy and robustness in recognizing the fine features of rice leaf diseases. Therefore, the application of this model to the intelligent identification of rice leaf disease can significantly improve the accuracy and reduce the misjudgment rate.

Why it matches plant phenotyping methodsイネ葉の病害状態を画像から識別する改良Faster-RCNNを開発・評価しており、植物病害表現型の取得手法が研究の中心である。

abstractan improved Faster-RCNN (Faster-RCNN-Pro) model is proposed to overcome these issues.

YOLOv11n-DualPC-Lite: a lightweight, high-precision real-time detection model for maize leaf diseases. →

To address the challenge of balancing model lightweight and detection accuracy in maize leaf disease detection, as well as the limitations of edge device deployment resources, we propose an enhanced target detection model, YOLOv11n-DualPC-Lite.Firstly, the C2fDualPConv module was designed, integrating PartialConv to replace some C3k2 modules in the backbone and neck networks. This approach enhances feature representation while reducing the number of parameters. Secondly, the Slim-Neck architecture is introduced in the neck network. To improve accuracy without increasing the number of parameters, the VoVGSCSPC_SimAm module enables the new Slim-Neck structure to reduce parameters while strengthening feature representation. Finally, an EfficientHead detection head is introduced that uses an inverted bottleneck MBConv module to improve performance. This significantly reduces computational load while efficiently extracting features. This study constructed a maize leaf disease dataset integrating a publicly available Kaggle dataset and a field-collected dataset from Anhui Science and Technology University's experimental plots. The dataset includes four categories: Blight, Common_Rust, Gray_Leaf_Spot, and Health. Through techniques such as rotation and gamma correction, the dataset was expanded from 3,876 to 5,165 images for model training and performance validation. Test results show this improved model performs better than other popular lightweight models overall, with a mAP50 score of 90.9%. Meanwhile, the model has only 2.13 million parameters; its computational complexity is reduced to 4.55 G, and the model size is 4.41 MB. Compared with the original YOLOv11n, its mAP50 is 1.9% higher, while the number of parameters is down by 17.8%, computational complexity is cut by 29.3%, and file size is reduced by 15.7%. When run on a Raspberry Pi 5, the model's detection speed reaches 2.3 FPS, an increase of 27.8%. This model achieves a good balance between detection accuracy and lightweight performance for maize leaf diseases, providing an efficient and practical method for real-time crop disease monitoring.

Why it matches plant phenotyping methodsトウモロコシ葉の病徴を画像から検出・分類する軽量モデルを開発し、データセット構築、性能比較、エッジデバイス検証まで行っており、植物病害状態の画像ベース表現型取得が中心である。

abstractwe propose an enhanced target detection model, YOLOv11n-DualPC-Lite

LeafFusionNet: a hybrid deep learning approach for robust plant disease detection →

Introduction Crop diseases have to be diagnosed early to save crop yields and food safety. Commonly used conventional practices are time-consuming and likely to involve human error. Automatic plant disease detection is a highly efficient technology that has been brought into existence by the development of deep learning and computer vision technology and can effectively detect the signs of diseases in plant species. Sustainable agriculture and early intervention depend on accurate and interpretable detection of plant diseases. Methods This study introduces a hybrid model based on deep learning techniques that effectively identifies and categorizes leaf diseases. The proposed model, LeafFusionNet, incorporates Convolutional Neural Network (CNN) and Vision Transformer (ViT) with an efficient attention module, LeafTAM (Leaf Texture Attention Module), to effectively capture both global and local information. This architecture is enhanced by the addition of a Gabor filter layer before the CNN-ViT fusion, hence augmenting the model’s ability to extract physiologically relevant texture features. Results and discussion The model was trained and validated on the Plant Village dataset. This model attained an accuracy of 99.33%, 99% precision, recall, and F1 score, demonstrating strong generalization on new data when compared with the state-of-the-art models. This proposed hybrid model can be utilized to develop a strong agricultural diagnostics system. These results point to the possibility of designing a powerful, interpretable system by utilizing transformer-based vision modules, Gabor filters, and explainability systems such as Grad-CAM.

Why it matches plant phenotyping methods植物葉の病徴を画像から検出・分類する深層学習手法を開発し、PlantVillageデータセットで訓練・検証しており、植物病害状態の表現型取得が研究の中心です。

abstractThis study introduces a hybrid model based on deep learning techniques that effectively identifies and categorizes leaf diseases.

AutoSiQ: a curated haploid Arabidopsis thaliana inflorescence dataset with a fine-grained silique ontology and a deep learning application for haploid fertility quantification. →

Doubled haploid (DH) technology can fast-track crop breeding. Haploid induction yields haploids with only one set of genomes, which are usually sterile. Haploid fertility (HF) is the ability of haploid plants to set seed, and it is a critical bottleneck in DH pipelines. Genetic mechanisms to restore HF hold immense potential in DH crop breeding, yet its phenotyping remains manual, destructive, and inconsistent. While recent advances in imaging and machine learning have improved throughput for general plant traits, no curated image dataset exists for Arabidopsis thaliana that explicitly represents HF. Here, we present AutoSiQ, a dataset and baseline deep learning pipeline for automated HF quantification. AutoSiQ includes high-resolution scanned inflorescences annotated with a seven-class ontology encompassing green siliques, green fertile siliques, mature siliques, fertile siliques, cracked fertile siliques, cracked siliques, and flowers. This multi-class annotation scheme preserves biologically meaningful information beyond binary fertile/non-fertile distinctions, enabling reliable fertility estimation and future phenotyping applications. We release baseline object detection models (YOLOv5), trained using the AutoSiQ dataset, and evaluate their performance across confidence thresholds. Model predictions strongly correlate with manual counts, achieving R² up to 0.94 for total silique number estimation. We further demonstrate AutoSiQ's utility for automated haploid fertility rate (HFR) estimation and genotype discrimination between two contrasting genotypes (WT and bmf2 mutant). A longitudinal analysis identifies ~60 days after sowing (DAS) as the optimal harvest time for maximizing mature silique counts by balancing between the number of immature buds and silique shattering. By releasing both the dataset and baseline code, AutoSiQ provides a reproducible and extensible foundation for high-throughput fertility phenotyping in haploid Arabidopsis .

Why it matches plant phenotyping methodsハプロイド稔性を画像から定量するデータセットと深層学習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 AutoSiQ, a dataset and baseline deep learning pipeline for automated HF quantification.

Significant increase in root exudation of 2'-deoxymugineic acid (DMA) as a response to zinc deficiency in rice →

1 Summary Zinc (Zn) deficiency limits rice productivity and poses a risk to human health, particularly in populations reliant on rice-based diets. Although rice germplasm exhibits wide variation in Zn-deficiency tolerance, the underlying physiological mechanisms remain poorly resolved. Evidence across the literature for Zn-deficiency–induced secretion of 2′-deoxymugineic acid (DMA) is inconsistent. This study clarifies the role of DMA secretion as a Zn-deficiency stress response. We developed and validated a sensitive LC–ESI–Q–TOF–MS method for selective detection of DMA in rice root exudates. Five rice genotypes with contrasting Zn-deficiency tolerance were grown hydroponically and DMA secretion measured. Zn-deficiency increased DMA exudation across all genotypes, with sensitive genotypes also showing higher secretion compared with control, supporting DMA’s role as a general response to Zn stress rather than being restricted to efficient genotypes. Fold-change responses exceeded previous studies, likely due to more severe stress exposure. Our results confirm that DMA secretion is induced under Zn-deficiency in rice as part of the micronutrient stress response. However, the lack of increased Zn uptake indicates that additional tolerance mechanisms are involved. These findings reconcile inconsistencies in the literature and position DMA secretion as an important, but not exclusive, component of Zn-deficiency adaptation in rice.

Why it matches plant phenotyping methodsイネ根滲出液中のDMAを選択的に検出するLC–MS法を開発・検証し、亜鉛欠乏応答という植物生理状態を測定しているため、化学分析が単なる付随測定ではなく中心的な方法貢献である。

abstractWe developed and validated a sensitive LC–ESI–Q–TOF–MS method for selective detection of DMA in rice root exudates.

InspectGaussian: Large-scale coarse-to-fine Gaussian reconstruction for orchard inspection robots →

Efficient large-scale 3D reconstruction of orchard environments is essential for robotic inspection and precision agriculture, yet existing methods struggle with unstructured scenes, variable illumination, and computational bottlenecks. We propose InspectGaussian, a coarse-to-fine Gaussian reconstruction framework tailored for orchard inspection robots. The pipeline integrates an RGB-D-based data acquisition strategy using ORB-SLAM3, which is enhanced by a dense mapping module for robust large-scale pose estimation and point cloud generation. A divide-and-conquer strategy is then employed: individual plant views are extracted via a YOLO-World-based detection and 3D matching algorithm, followed by plant-specific reconstruction using an improved 3D Gaussian Splatting (3DGS) method incorporating depth regularization and region-aware refinement. Experimental results in citrus orchards demonstrate that InspectGaussian achieves 96% average precision and 93% recall in plant view extraction, while surpassing state-of-the-art methods in reconstruction fidelity (31.226 PSNR, 0.915 SSIM, 0.067 LPIPS) and point cloud accuracy (7 mm error). These results confirm its effectiveness in capturing fine structural and textural details while maintaining scalability and efficiency. This framework provides a practical solution for high-throughput, in-field plant phenotyping and lays the foundation for intelligent orchard monitoring and management.

Why it matches plant phenotyping methods植物個体の3D再構成とRGB-D・検出・Gaussian Splattingを統合した手法開発であり、植物の構造的形質取得を目的とするため、フェノタイピング手法が中心である。

abstractWe propose InspectGaussian, a coarse-to-fine Gaussian reconstruction framework tailored for orchard inspection robots.

Tomato Multi-Angle Multi-Pose Dataset for Fine-Grained Phenotyping. →

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.

A field-acquired RGB-Depth image dataset for computer vision-based baby broccoli detection and size estimation under varying illumination conditions. →

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.

Classification of rice plant diseases using efficient DenseNet121 →

Abstract Agriculture and global food security are critically dependent on accurate and timely identification of plant diseases and pests. Traditional approaches to disease identification rely heavily on visual inspection and expert knowledge, which frequently lack the accuracy, speed, and scalability needed to address growing agricultural challenges. Early and precise disease detection enables proactive interventions that can prevent widespread crop damage and reduce excessive pesticide use, thereby supporting sustainable agricultural practices. Artificial intelligence, particularly deep learning methods, has emerged as a transformative solution for automated plant disease diagnosis. Convolutional neural networks (CNNs) have demonstrated remarkable capabilities in image classification tasks, evolving from individual architectures to sophisticated ensembles and transferring learning models. However, existing CNN-based research on rice disease identification has typically focused on a limited number of disease classes, restricting their practical applicability in real-world agricultural settings. This study addresses these limitations by implementing DenseNet121, an advanced CNN architecture known for its efficient feature reuse and gradient flow, for comprehensive rice disease classification. We utilized a dataset comprising seven of the most common rice diseases, significantly expanding the scope beyond previous studies. The model employs transfer learning with pre-trained ImageNet weights and is optimized using the Adam optimizer with carefully tuned hyperparameters. The experimental evaluation on an independent test set demonstrates that our proposed model achieves an overall accuracy of 97.9%, with individual disease classification accuracy ranging from 94% to 99.67%. The model exhibits balanced performance across multiple metrics, including precision (96.2%), recall (97.97%), and F1-score (97%), confirming its robustness and generalizability. These results establish DenseNet121 as a highly effective framework for automated rice disease diagnosis, offering a practical tool for enhancing agricultural productivity and food security.

Why it matches plant phenotyping methodsイネ葉の画像から病害状態を分類する深層学習手法が研究の中心であり、独立テストセットによる性能評価も実施しているため、植物病害フェノタイピング手法として収載する。

abstractThis study addresses these limitations by implementing DenseNet121, an advanced CNN architecture known for its efficient feature reuse and gradient flow, for comprehensive rice disease classification.

A Comparative Study of Deep Transfer Learning Architectures for Multi-Class Plant Leaf Disease Detection →

Plant Leaf Diseases represent a significant risk to global agricultural production. Crops ranging from Peppers and Tomatoes to Potatoes are affected by these diseases. Traditional methods of identifying leaf diseases based primarily on visual inspection have historically been slow and relatively inaccurate. A deep learning-based solution for the automated identification of plant leaf diseases utilizes Convolutional Neural Networks (CNNs) as the primary methodology. To identify leaf images into 15 disease categories, both pre-trained models such as VGG16, EfficientNetB3, Inception V3 and a custom CNN model were used. Using pre-trained models to allow for the use of Transfer Learning helps to mitigate some of the issues associated with computational resource limitations and data limitations in providing faster convergence rates and higher accuracy when compared to training a model from scratch. The Plant Village image collection which contains over 20,000 images of different plant leaf diseases was utilized for training and testing purposes. Each model's performance was evaluated based on its accuracy, loss and generalization capabilities. Additionally, each model was fine-tuned through hyperparameter optimization. As a result, the model that achieved the highest validation accuracy rate of 95% was the EfficientNetB3 model while the second highest accuracy rate was achieved by the Inception V3 model at 92%. This methodology provides an excellent answer to addressing early disease detection, enabling farmers to take the necessary actions quickly to reduce their losses and maximize their harvest.

Why it matches plant phenotyping methods植物葉画像から病害状態を推定する深層学習手法を比較・評価しており、病害フェノタイピングと手法検証が研究の中心です。

abstractA deep learning-based solution for the automated identification of plant leaf diseases utilizes Convolutional Neural Networks (CNNs) as the primary methodology.

Precise tea leaf disease detection using UAV low-altitude remote sensing and optimized YOLO11 model. →

Tea leaf diseases seriously affect its yield and quality, and consequently there is an urgent need for intelligent detection methods with high precision and edge deployment capabilities. To address low detection accuracy in complex backgrounds, overfitting due to limited data, and redundant parameters for existing methods, this paper proposes an improved lightweight detection model FCHE-YOLO based on the YOLO11, which aims to achieve rapid and accurate identification of tea leaf disease combining low altitude remote sensing with unmanned aerial vehicle (UAV). The model has made three key optimizations in the structure: Introduce the self-developed lightweight backbone module FC_C3K2, which significantly reduces computation and parameter count while enhancing the robustness of the model to complex scenarios; construct an efficient feature fusion structure HSFPN, optimizing multi-scale information integration and compressing model volume; design the detection head Efficient Head, integrating group convolution and lightweight attention mechanism to improve detection accuracy and suppress overfitting. The experimental results from the self built tea gardens show that the FCHE-YOLO improves the average accuracy (mAP) from 94.1% to 98.1% compared to the benchmark model YOLO11, with an improvement of 4.0 percentage points. Meanwhile, the inference speed of the model increases from 43.3 FPS to 47.5 FPS, with an increase of 9.0%, meeting the real-time detection requirements. More importantly, by network structure optimization, the model's computational complexity is significantly reduced: The floating-point operations per second (FLOPs) decreases from 6.4 G to 4.2 G, with a decrease of 34.3%, and the parameter count decreases from 2.59 M to 1.46 M, with the compression rate reaching 38.9%, which makes the model more suitable for deployment on resource-constrained UAV edge devices. The final test show that the FCHE-YOLO significantly reduces the missed-detection rate, owns better detection accuracy and deployment practicality, and is suitable for real-time monitoring scenarios of tea leaf diseases with UAVs.

Why it matches plant phenotyping methods茶葉の病害状態をUAV画像から検出する軽量深層学習手法を開発・評価しており、植物病害表現型の取得が中心的な技術貢献である。

abstractthis paper proposes an improved lightweight detection model FCHE-YOLO based on the YOLO11, which aims to achieve rapid and accurate identification of tea leaf disease combining low altitude remote sensing with unmanned aerial vehicle (UAV).

A hybrid deep learning framework using convolutional and transformer models for robust plant disease classification →

Abstract Plant diseases continue to pose a significant threat to worldwide food security, resulting in notable yield reductions and economic consequences. Automated disease diagnosis through machine learning has arisen as a potential solution; nevertheless, current methods frequently have difficulty in capturing both detailed local attributes and overarching contextual patterns found in plant leaf images. This study presents a thorough comparative examination of conventional and deep learning methods—such as Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), YOLO, Support Vector Machines (SVMs), and Random Forests—for the classification of multi-class plant diseases. To overcome the constraints of individual CNN and transformer models, a new hybrid framework that integrates EfficientNet-B7 for strong spatial feature extraction with a Vision Transformer (ViT-B16) for comprehensive contextual modeling is suggested. The system is assessed on an extensive dataset consisting of 21,534 images covering 38 classes of plant diseases and healthy specimens. Experimental findings show that the suggested hybrid model reaches an accuracy of 98.13%, surpassing standalone CNN baselines and other rival models, while consistently achieving high precision, recall, and F1-scores for all classes. The results emphasize the success of combining convolutional and transformer-based models for scalable and precise plant disease detection, aiding the creation of smart decision-support systems for precision farming.

Why it matches plant phenotyping methods植物葉画像から病害状態を分類する新規ハイブリッド画像解析手法を提案し、複数手法との比較評価と大規模データセットでの検証を行っており、フェノタイピング手法が中心である。

abstractAutomated disease diagnosis through machine learning has arisen as a potential solution

Leaf-DETR: Progressive adaptive network with lower matching cost for dense leaves detection. →

Leaves are central indicators of photosynthesis and plant growth status, and their precise monitoring is crucial for smart agriculture. Dense leaf detection, as a foundation for leaf morphology analysis, must address challenges such as occlusion and overlap, directly enabling key tasks including phenotypic trait extraction, disease identification, and yield estimation. Leaves are the most important plant organs, and monitoring leaves is a crucial aspect of crop surveillance. Dense leaf detection plays an important role as a fundamental technology for leaf monitoring. Existing dense leaf detection methods rely on traditional modular detectors and generic feature extraction, lacking designs tailored to real-world dense leaf scenarios. The methods for dense leaf detection generally use traditional modular detectors and general feature extraction techniques, without designing methods specifically for dense leaves in reality. In detail, in complex field scenarios, it still faces challenges like incomplete individual feature extraction due to high leaf overlap and difficult network convergence caused by excessive leaf density. To this end, we propose the Leaf-DETR framework, which effectively addresses these challenges through the Progressive Feature Fusion Pyramid Network (P-FPN) and the Crowded Query Refinement Strategy (CQR). First, we construct the largest dense leaf detection dataset to date, containing 1696 images and 85,375 annotation boxes. Second, P-FPN alleviates the feature confusion problem of overlapping leaves through the multi-stage fusion of features and the Adaptive Feature Aggregation module (AFA), enhancing the interaction between low-level details and high-level semantics. Third, the CQR strategy significantly reduces the matching cost of crowded candidate boxes and improves the network convergence efficiency by culling a crowded query method and introducing a one-to-many matching mechanism. Finally, experimental results show that Leaf-DETR improves mAP@50 by 1% and AR@300 by 1.4% over the baseline model on our self-constructed dataset, outperforming existing detection methods. Furthermore, the model exhibits extremely fast training convergence and demonstrates strong generalization capability on both field-collected monitoring images and other staple crops, fully highlighting its practical value in complex agricultural scenarios. Finally, experiments show that Leaf-DETR outperforms existing detection methods on the self-built dataset and demonstrates good performance generalization in monitoring collected images, as well as for other staple food crops, which verifies its practicality in complex agricultural scenarios. The code and detailed information are available at http://leafdetr.samlab.cn.

Why it matches plant phenotyping methods葉の密集検出モデルとデータセットを開発・評価し、葉形態などの表現型抽出を可能にする画像ベース手法が研究の中心であるため。

abstractDense leaf detection, as a foundation for leaf morphology analysis, must address challenges such as occlusion and overlap, directly enabling key tasks including phenotypic trait extraction, disease identification, and yield estimation.

Pepper-4D: Spatiotemporal 3D Pepper Crop Dataset for Phenotyping →

Pepper (Capsicum annuum) is a globally significant horticultural crop cultivated for its culinary, medicinal, and economic value. Traditional approaches for boosting the agricultural production of pepper, notably, expanding farmland, have become increasingly unsustainable. Recent advancements in artificial intelligence and 3D computer vision have started to transform crop cultivation and phenotyping, which has shed new light on increasing production by advanced breeding. However, currently, the field still lacks 3D pepper data that contains enough detail for organ-level analysis. Therefore, we propose Pepper-4D, a new, high-precision 4D point cloud dataset that records both the spatial structure and temporal development of pepper plants across various continuous growth stages. Our dataset is divided into three subsets, including a total of 916 individual point clouds from 29 indoor-cultivated pepper plant samples. Our dataset provides manual annotations at both the plant-level and organ-level, supporting phenotyping tasks such as pepper growth status classification, organ semantic segmentation, organ instance segmentation, organ growth tracking, new organ detection, and even the generation of synthetic 3D pepper plants.

Why it matches plant phenotyping methods植物の器官レベル表現型解析を支援する4D点群データセットを構築し、成長状態分類・器官分割・追跡などを可能にする研究であり、フェノタイピング用データ基盤が中心です。

abstractPepper-4D, a new, high-precision 4D point cloud dataset that records both the spatial structure and temporal development of pepper plants across various continuous growth stages.

Lightweight scalable deep learning framework for real time detection of potato leaf diseases. →

Potato leaf diseases, if left undetected, threaten food security in agricultural economies and cause substantial crop losses. To address this critical challenge, we developed an AI-based system called the Enhanced Single Shot Multibox Detector(EF-SSD), a variant of SSD that integrates multiscale feature fusion and Squeeze-and-Excitation attention to improve fine-grained lesion detection. The enhanced model processes high-resolution images (512×512 pixels) and analyzes leaves at ten magnification levels, enabling it to identify even minor signs of infection. The inclusion of Squeeze-and-Excitation filters allows the system to focus more effectively on characteristic disease patterns, increasing detection precision. After scanning the leaves, the system applies advanced image processing techniques to localize disease regions and assess their severity. We evaluated EF-SSD using 2,500 labeled potato leaf images representing healthy plants and cases of early and late blight. The proposed model achieved a mean Average Precision (mAP) of 97% at 0.5 IoU, an F1-score of 95%, and an Intersection over Union (IoU) of 89%, outperforming advanced detectors such as YOLOv5, YOLOv8, RetinaNet, and Faster R-CNN across all metrics. It also delivers real-time inference at 47 FPS, confirming its suitability for on-field deployment. An ablation study further demonstrates the effectiveness of SE blocks and extended feature hierarchies in enhancing detection accuracy. These outcomes highlight EF-SSD’s potential as a reliable, efficient, and scalable tool for smart agriculture and early crop disease management.

Why it matches plant phenotyping methodsジャガイモ葉の病斑を画像から検出・局在化し、病害の重症度を評価する深層学習手法を開発・検証しており、植物表現型取得が中心である。

abstractwe developed an AI-based system called the Enhanced Single Shot Multibox Detector(EF-SSD)

Dual-encoder Variational Autoencoder for Detection and Classification of Plant Leaf Diseases →

Background: Plant disease detection remains a major challenge in agriculture, with direct implications for improving crop productivity and ensuring food security. Seasonal variation significantly influences plant characteristics, making the classification of plant leaves by season-specifically summer and winter-important for optimizing disease detection and management strategies. Methods: In this study, a plant leaf disease detection dataset was developed and categorized based on seasonal conditions. The dataset includes 47 classes representing summer crops and 16 classes for winter crops. To classify plant leaf diseases effectively, we propose a novel dual-encoder Variational Autoencoder (VAE) model that integrates ResNet and VGGNet as parallel encoders. These encoders extract complementary feature maps from the seasonal datasets, which are then concatenated to improve classification accuracy. Result: Experimental evaluation demonstrates the robustness and accuracy of the proposed approach. The dual-encoder VAE achieved a classification accuracy of 98.86% on the summer dataset and 97.53% on the winter dataset, highlighting the model’s ability to generalize effectively across seasonal variations in plant leaf disease detection.

Why it matches plant phenotyping methods植物葉の病徴を画像から検出・分類する新規VAE手法を開発し、季節別データセットで精度評価しているため、病害状態の表現型取得が中心である。

abstractTo classify plant leaf diseases effectively, we propose a novel dual-encoder Variational Autoencoder (VAE) model that integrates ResNet and VGGNet as parallel encoders.

Deep learning enables quantitative subcellular analysis of plant-microbe interfaces →

Specialized host-microbe interfaces are central to cellular interactions in plants. Intracellular structures such as haustoria formed by filamentous pathogens mediate nutrient exchange and effector delivery to host cells. Despite their biological importance, the lack of quantitative frameworks has largely confined the study of these interfaces to qualitative observations, limiting our ability to compare infection strategies, cellular responses, and spatial organization across cells and tissues. Here, we present HFinder , a deep learning-based framework for automated detection, segmentation, and quantitative analysis of plant-microbe interfaces in confocal images. Using an object-centric deep learning approach, HFinder enables robust identification of haustoria, microbial hyphae, and host organelles across diverse imaging conditions and pathosystems. We demonstrate that this framework supports quantitative analyses of subcellular processes at host-microbe interfaces, including effector secretion, perturbation of host cellular processes, and immune receptor accumulation at haustoria. HFinder provides a practical and scalable solution for the systematic digitalization of plant infection imaging data and establishes a general framework for quantitative studies of cellular dynamics at host-microbe contact zones.

Why it matches plant phenotyping methods植物と微生物の界面を共焦点画像から自動検出・分割し、ハウストリア等を定量解析する深層学習手法が中心であり、植物感染状態の画像ベース表現型解析に該当する。

abstractwe present HFinder , a deep learning-based framework for automated detection, segmentation, and quantitative analysis of plant-microbe interfaces in confocal images.

Branch architecture reconstruction and phenotypic trait analysis of poplar trees using low-cost UAV LiDAR point clouds. →

Poplars are essential to China's forestry, contributing to timber production, ecological restoration, and shelterbelt construction. Branch architecture critically influences tree growth, demanding scalable solutions beyond manual methods to assess phenotypic variation in large-scale poplar breeding programs. Unmanned aerial vehicle light detection and ranging (UAV LiDAR) provides an efficient alternative; however, existing methods focus on conifers, leaving a gap in approaches for the more complex morphology of poplar branches. This study proposes a poplar branch reconstruction algorithm utilizing material transport flux and object-level geometric features from low-cost UAV LiDAR data. First, a voxel-based near-centroid method is used to extract skeleton points from tree point clouds. Next, a material transport flux model identifies individual branches, and geometric features of transport paths, including path length and curvature, are used to reconstruct each branch. Finally, branch parameters are estimated based on reconstructed branches. Data from a 5-ha plot were collected using the DJI Zenmuse L1 UAV LiDAR at the Shishou National Poplar Breeding Station, Hubei Province, China. Results demonstrate the proposed algorithm achieves high accuracy in first-order branch identification (F1-score = 1), with second-order branches having an average F1-score of 0.69. Branch length estimation demonstrates an RMSE of 0.47 m, while branch angles show an RMSE of 7.06°. The study also reveals structural variability in branch traits, with the highest variability observed in the second-order branch length (coefficient of variation = 29.68%), and a moderate positive correlation between first- and second-order branch lengths (correlation coefficient = 0.34), providing insights into tree growth patterns. This approach offers a framework for high-throughput phenotyping, which provides an efficient solution towrads advanced tree breeding using UAV LiDAR.

Why it matches plant phenotyping methodsUAV LiDARによるポプラの枝構造再構成と枝長・枝角度などの形質推定アルゴリズムを開発し、精度検証まで行っており、フェノタイピング手法が研究の中心である。

abstractThis study proposes a poplar branch reconstruction algorithm utilizing material transport flux and object-level geometric features from low-cost UAV LiDAR data.

A digital twin-driven deep learning framework for online quality inspection in tobacco transplanting. →

Tobacco transplanting quality inspection is crucial for tobacco production, as it directly affects crop yield and quality of tobacco leaves. Accurate transplanting status detection and assessment provide essential support for replanting decisions and transplanting machine optimization. Traditional methods rely on manual inspection, which suffer from high cost, low efficiency, and unstable results. To tackle the aforementioned issues, this paper proposes a Deep Learning and Digital Twin driven Online Quality Inspection Method for Tobacco Transplanting, which consists of four core modules: Transplanting Status Detection, Multi-sensor Data Fusion, Digital Twin Visualization, and Operational Optimization Feedback. This paper proposes a lightweight improved YAN-YOLO11 algorithm capable of assessing normal, exposed-root, and buried seedlings. By fusing GNSS positioning data with visual detection results, the system estimates in-row spacing and performs status assessment for missed planting and double planting. The system establishes a virtual-real interactive closed-loop of "collection-detection-mapping-feedback" via the digital twin. By visualizing operational status in real-time and generating replanting path suggestions, it provides guidance for operation management and significantly improves inspection efficiency. Field experiments demonstrate that, compared with YOLO11n, YAN-YOLO11 improves precision and recall by 2.4% and 2.5%, respectively; mAP@50 increased by 3% to 80.9% ± 1.4%, and mAP@0.5:0.95 increased by 5.8% to 54.2% ± 1.0%, while significantly reducing model complexity. The system achieves a real-time performance of 30 FPS in the field, with an overall recognition accuracy of 90.74%, meeting practical application requirements. This study effectively enhances the digitalization, automation, and refined management of tobacco transplanting operations, providing a theoretical foundation and practical solution for the intelligent transformation of transplanting machinery and precision crop management.

Why it matches plant phenotyping methodsタバコ苗の植栽状態、株間、欠株・二株植えを画像とセンサーで推定する手法を開発・検証しており、植物状態の取得が研究の中心である。

abstractThis paper proposes a lightweight improved YAN-YOLO11 algorithm capable of assessing normal, exposed-root, and buried seedlings.

Unsupervised Tree Detection from UAV Imagery and 3D Point Clouds via Distance Transform-Based Circle Estimation and AIC Optimization →

This work proposes a novel tree detection methodology, named DTCD (Distance Transform Circle Detection), based on a fast circle detection method via Distance Transform and Akaike Information Criterion (AIC) optimization. More specifically, a visible-band vegetation index (RGBVI) is calculated to enhance canopy regions, followed by morphological filtering to delineate individual tree crowns. The Euclidean Distance Transform is then applied, and the local maxima of the smoothed distance map are extracted as candidate tree locations. The final detections are iteratively refined using the AIC to optimize the number of trees with respect to canopy coverage efficiency. Additionally, this work introduces DTCD-PC, a modified algorithm tailored for point clouds, which significantly enhances detection accuracy in complex environments. This work makes a significant contribution to tree detection in the following ways: (1) by creating a tree detection framework entirely based on an unsupervised technique, which outperforms state-of-the-art unsupervised and supervised tree detection methods; (2) by introducing a new urban dataset, named AgiosNikolaos-3, that consists of orthomosaics and photogrammetrically reconstructed 3D point clouds, allowing the assessment of the proposed method in complex urban environments. The proposed DTCD approach was evaluated on the Acacia-6 dataset, consisting of UAV images of six-month-old Acacia trees in Southeast Asia, demonstrating superior detection performance compared to existing state-of-the-art techniques, both unsupervised and supervised. Additional experiments were conducted in the custom-developed Urban Dataset, confirming the robustness and generalizability of the DTCD-PC method in heterogeneous environments.

Why it matches plant phenotyping methodsUAV画像・3D点群から個体樹冠を抽出する新規手法を開発し、複数データセットで精度・頑健性を評価しているため、植物形態の取得・抽出が中心である。

abstractThis work proposes a novel tree detection methodology, named DTCD (Distance Transform Circle Detection), based on a fast circle detection method via Distance Transform and Akaike Information Criterion (AIC) optimization.

A wavelet-based frequency-domain approach for accurate multi-crop disease detection. →

Timely localization and diagnosis of crop lesions are critical for disease control and reducing pesticide use. However, in-field lesions often resemble leaf textures, vary widely in scale, and suffer from lighting and shadow interference-making simultaneous high accuracy and lightweight inference challenging. We propose WGA-YOLO, a lightweight YOLO variant for crop disease recognition. Central to our design is Wavelet Channel Recalibration (WCR), a DWT-based downsampling module: discrete wavelet transform naturally provides multi-resolution, time-frequency localized representations that explicitly separate low-frequency approximations from high-frequency edge/texture details. WCR fuses high- and low-frequency components and enhances feature representation through their frequency-domain complementarity, thereby preserving semantic and fine texture information during resolution reduction with negligible extra cost. We also introduce PS-C2f, which integrates Pinwheel-shaped convolutions into C2f to better capture tiny lesion details via multi-directional, irregular kernels, and replace SPPF with Dynamic Group Attention Pooling (DGAP) for efficient multi-scale context aggregation. On our PlantDoc_boost dataset, WGA-YOLO improves over YOLOv8n by 3.02 and 2.85% points, while reducing parameters and FLOPs by ~ 0.18 M and ~ 0.3G, demonstrating improved inference efficiency and deployment friendliness while maintaining strong detection performance in field scenarios.

Why it matches plant phenotyping methods植物葉の病斑を画像から検出・診断するYOLO改良手法の開発が中心であり、病害状態の画像ベース表現型計測に該当する。

abstractWe propose WGA-YOLO, a lightweight YOLO variant for crop disease recognition.

Camera-based bi-axial measurement of weak forces generated by freely moving plant organs →

Growing plants are remarkable at negotiating obstacles in their unstructured and changing environments. Measuring the mechanical interactions of growing plants with surrounding objects is a critical step towards deciphering thigmotropic responses underpinning complex growth strategies. Yet, available force measurement systems have limited capacity to capture weak forces in freely moving plant organs-such as the forces applied by a growing shoot pushing at an obstacle. We developed a measurement system based on the deflection of a pendulum by a freely moving shoot. Unlike many force measurement systems, the organ is not tethered to the device. Moreover, force is measured along two axes, as opposed to one axis in commonly used methods. Orthogonal cameras track the 3D position of the rod and shoot, yielding the rod deflection angle and, using a mechanical torque equilibrium equation, allowing extraction of the force applied by the plant over time. This system is relevant for measuring weak forces in macro-sized systems (e.g. growth or turgor pressures), and the force detection range can be tuned by altering rod mass and length. We demonstrate the system with Phaseolus vulgaris shoots, measuring the forces they apply on a candidate support during inherent circumnutation movements, prior to twining. Such measurements lay the foundations for deciphering how climbing plants assess whether to twine or not- an open question since Darwin's first observations.

Why it matches plant phenotyping methods自由に動く植物器官が発生する微弱な力を、カメラ追跡と力抽出により定量する測定システムを開発・実証しており、植物表現型の取得方法が研究の中心である。

abstractWe developed a measurement system based on the deflection of a pendulum by a freely moving shoot.

LeafAI: Interpretable plant disease detection for edge computing →

In real-world agriculture, healthy plant leaves are significantly more common than diseased ones. This natural class imbalance presents challenges in automated plant disease detection, as analyzing each leaf with computationally intensive deep-learning models is problematic, leading to inefficiency and increased resource consumption. To tackle this challenge and promote sustainable AI solutions, this study presents an iterative, hybrid AI approach that boosts computational efficiency, interpretability, and scalability for real-time disease detection. This hybrid system operates in two stages: first, a lightweight traditional machine learning classifier performs binary classification to quickly separate and exclude healthy leaves, followed by a deep learning model (ResNet, DenseNet, MobileNet, and EfficientNet) that classifies the specific disease in the smaller group of diseased leaves. This two-stage method minimizes computational load while maintaining high classification accuracy. Additionally, this study uses Explainable AI (XAI) methods, particularly Gradient-weighted Class Activation Mapping (Grad-CAM), to generate heatmaps. These heatmaps highlight the image areas that most significantly influence the model’s predictions, thereby improving transparency and refining the feature extraction process. The proposed hybrid model, comprising Logistic Regression and Mobilenetv3, offers up to 77.6% faster inference than conventional deep learning models with only about 3% accuracy loss. For a large-scale test of 1,227 images on an entry-level laptop, the hybrid model reduced the total inference time from 4,548 seconds to just 1,010.13 seconds, with minimal CPU load. By addressing class imbalance, optimizing inference efficiency, and incorporating explainable AI, this work contributes a scalable, sustainable, and trustworthy solution for plant disease detection in precision agriculture.

Why it matches plant phenotyping methods葉画像から植物病害状態を推定する二段階AI手法とXAIを開発・評価しており、病害の表現型取得が研究の中心である。

abstractThis two-stage method minimizes computational load while maintaining high classification accuracy.

An Ensemble Convolutional Neural Network Framework for Automated Mango Leaf Disease Detection →

Abstract Mango diseases and pest infestations represent a major challenge to agricultural productivity, making early and accurate diagnosis crucial for reducing crop losses. This study presents a security-preserving ensemble convolutional neural network (CNN) framework for the automated identification and classification of mango leaf diseases using image-based analysis. The proposed system is designed to work with images captured under real field conditions, ensuring its suitability for practical agricultural applications. The dataset includes mango leaf images affected by various diseases and pests such as Gall Midge, Powdery Mildew, Sooty Mould, Die Back, Cutting Weevil, and Anthracnose, each characterized by distinct visual symptoms including discoloration, necrotic spots, fungal growth, leaf deformation, and edge damage. Traditional manual diagnosis of these conditions is often time-consuming, labor-intensive, and susceptible to human error. To overcome these limitations, the proposed framework employs an ensemble of transfer-learning-based CNN models to extract meaningful features related to texture, color distribution, shape, and lesion patterns. A security-preserving learning mechanism is integrated to ensure the safe handling of agricultural image data, minimizing data exposure risks while maintaining high model performance. Additionally, data augmentation techniques are utilized to improve model robustness, reduce overfitting, and address class imbalance commonly found in agricultural datasets. The system is capable of multi-class classification, reflecting real-world scenarios where multiple diseases may exhibit visually similar characteristics. Experimental results indicate that the ensemble CNN framework achieves high classification accuracy and demonstrates strong generalization across varying lighting conditions and complex backgrounds. By effectively capturing disease-specific visual features, the proposed approach enhances detection reliability in real-world field environments. Overall, this system offers a scalable, non-invasive, and security-aware solution for early mango leaf disease detection, contributing to precision agriculture and informed decision-making. The findings highlight the potential of deep learning and computer vision technologies in developing intelligent, secure, and efficient plant health monitoring systems.

Why it matches plant phenotyping methodsマンゴー葉の病徴を画像から分類するCNNフレームワークの開発が研究の中心であり、植物の病害状態を直接推定する画像ベース表現型解析に該当する。

abstractThis study presents a security-preserving ensemble convolutional neural network (CNN) framework for the automated identification and classification of mango leaf diseases using image-based analysis.

Chat Demeter: a multi-agent system for plant disease diagnosis integrating CNN-transformer models →

Plant diseases remain a significant challenge in global agricultural production. Achieving efficient and accurate disease detection is essential for reducing crop losses, controlling agricultural costs, and improving yields. As agriculture rapidly advances toward digitalization and intelligent transformation, the application of artificial intelligence technologies has become a key pathway to enhancing industrial competitiveness. In this study, Chat Demeter, a multi-agent system for plant disease diagnosis based on deep learning. The system captures real-time leaf images through camera devices. It employs a CNN-Transformer model to perform instance segmentation and object detection, thereby enabling automatic identification of diseased leaves and classification of disease types. To enhance interactivity and practical value, the system incorporates a natural language interface, allowing users to upload images and receive automated diagnostic results and treatment suggestions. Experimental results demonstrate that the system achieves an accuracy of 99.50% and an AUC o f 99.91% on the validation dataset, highlighting its superior performance. Overall, Chat Demeter provides an effective tool for crop health monitoring and disease intervention, while offering a feasible pathway and developmental direction for integrating and optimizing future agricultural multi-agent systems.

Why it matches plant phenotyping methods植物葉の画像から病葉をセグメンテーション・分類する診断システムが研究の中心であり、植物病害状態の画像ベース表現型計測に該当する。

abstractChat Demeter, a multi-agent system for plant disease diagnosis based on deep learning.

Eff-swin-hgso: attention-driven and optimized plant leaf disease diagnosis using efficientNetV2B0 and swin transformer with HGSO feature selection →

Abstract Agriculture is essential to human civilization, providing food and raw materials. Plant diseases significantly threaten agricultural productivity, making early and accurate detection essential. Despite Recent advances of deep learning in making automatic plant leaf diseases diagnosis systems, some of them depend on simple features fusion methods and lack an efficient method to exploit complementary information. Therefore, this paper proposes a system for diagnosing plant leaf diseases by fusing two powerful deep learning models: EfficientNetV2B0 and Swin Transformer via attention-based feature fusion that adaptively weights each model’s contribution with features. These models extract complementary features: EfficientNetV2B0 extracts fine-grained local features, and the Swin Transformer extracts global contextual information, producing highly and complementary expressive fused features. The high-dimensional fused features demand High-Performance Computing (HPC) resources for efficient parallel processing and accelerated training. Moreover, the Henry Gases Solubility Optimization (HGSO) metaheuristic is applied to select the most discriminative and related features. Unlike previous methods that diagnose diseases affecting only one plant, the proposed approach handles multiple plant species simultaneously, further increasing computational demand. Finally, RBF-kernel SVM is applied for a classification step. The system was implemented on a GPU-based high-performance computing environment using CUDA acceleration to enhance computational efficiency. Experimental evaluation on the PlantVillage benchmark dataset with seven classes achieved a high classification accuracy of 99.2%, outperforming other state-of-the-art methods. These results enhance the model’s practical capability for application in real-world agricultural decision-support systems.

Why it matches plant phenotyping methods葉画像から植物病害を診断する深層学習システムの開発とベンチマーク評価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。

abstractTherefore, this paper proposes a system for diagnosing plant leaf diseases by fusing two powerful deep learning models: EfficientNetV2B0 and Swin Transformer via attention-based feature fusion

Forest Surveying with Robotics and AI: SLAM-Based Mapping, Terrain-Aware Navigation, and Tree Parameter Estimation →

Forest surveying and inspection face significant challenges due to unstructured environments, variable terrain conditions, and the high costs of manual data collection. Although mobile robotics and artificial intelligence offer promising solutions, reliable autonomous navigation in forest, terrain-aware path planning, and tree parameter estimation remain open challenges. In this paper, we present the results of the AI4FOREST project, which addresses these issues through three main contributions. First, we develop an autonomous mobile robot, integrating SLAM-based navigation, 3D point cloud reconstruction, and a vision-based deep learning architecture to enable tree detection and diameter estimation. This system demonstrates the feasibility of generating a digital twin of forest while operating autonomously. Second, to overcome the limitations of classical navigation approaches in heterogeneous natural terrains, we introduce a machine learning-based surrogate model of wheel–soil interaction, trained on a large synthetic dataset derived from classical terramechanics. Compared to purely geometric planners, the proposed model enables realistic dynamics simulation and improves navigation robustness by accounting for terrain–vehicle interactions. Finally, we investigate the impact of point cloud density on the accuracy of forest parameter estimation, identifying the minimum sampling requirements needed to extract tree diameters and heights. This analysis provides support to balance sensor performance, robot speed, and operational costs. Overall, the AI4FOREST project advances the state of the art in autonomous forest monitoring by jointly addressing SLAM-based mapping, terrain-aware navigation, and tree parameter estimation.

Why it matches plant phenotyping methods自律ロボット、3D点群、深層学習を用いて樹木の直径・高さを推定する手法を開発し、点群密度による推定精度も評価しており、植物形質取得が中心である。

abstractwe develop an autonomous mobile robot, integrating SLAM-based navigation, 3D point cloud reconstruction, and a vision-based deep learning architecture to enable tree detection and diameter estimation.

Agentic AI for smart and sustainable precision agriculture. →

Introduction Ensuring smarter and more sustainable farming practices is a critical challenge in modern agriculture. Agentic Artificial Intelligence (AAI), combined with Precision Agriculture (PA) and Federated Learning (FL), has the potential to enhance decision-making, optimize resource utilization, and reduce environmental impact. Methods This study proposes an AAI based framework for precision agriculture that integrates distributed sensing devices, intelligent agents, and federated learning to enable real time monitoring and decision support at the farm level. A practical deployment architecture is outlined, detailing inter-device communication and localized intelligence. The proposed model is evaluated across two distinct datasets tomato disease classification and weed detection. The model is designed to have DenseNet121, MobileNetV2, EfficientDet-D0, and YOLOv8 as local models within a federated learning environment. Results The federated global model achieved an accuracy of 96.4%, outperforming individual client models, with DenseNet121 and MobileNetV2 attaining accuracies of 95.0% and 93.9%, respectively. For weed species detection, EfficientDet-D0 demonstrated superior performance, achieving an mAP@0.5 of 0.978, average precision of 0.865, and an F1-score of 0.961, compared to YOLOv8 with an mAP@0.5 of 0.956 and an F1-score of 0.935. Discussion The results confirm the feasibility and effectiveness of integrating AAI with federated learning for intelligent precision agriculture. A SWOT analysis highlights the strengths of the proposed approach, along with deployment challenges and constraints. Overall, this study establishes a roadmap for future research, emphasizing sustainable intelligent farming systems.

Why it matches plant phenotyping methods植物病害分類を含む連合学習・エージェント型AI基盤を提案し、データセット上で性能評価しているため、植物の病害状態を推定する計算的フェノタイピング手法が中心である。雑草検出も含まれるが、病害分類の技術評価が明示されている。

abstractThis study proposes an AAI based framework for precision agriculture that integrates distributed sensing devices, intelligent agents, and federated learning to enable real time monitoring and decision support at the farm level.

IoT-Integrated robotic system for automated plant disease detection and environmental monitoring →

Abstract Plant diseases pose a critical threat to global food security, agricultural sustainability, and farmer livelihoods, particularly in regions with limited access to advanced diagnostic technologies. Traditional methods of disease detection rely heavily on manual inspection, which is time-consuming, error-prone, and often results in delayed interventions. This paper presents a novel, solar-powered autonomous robotic system designed to detect plant diseases in real time using deep learning and IoT technologies. The proposed system integrates a high-resolution imaging unit, IoT-based environmental sensors, and an onboard processing module based on Raspberry Pi. Deep CNNs, trained on diverse datasets including PlantVillage, are used for accurate disease classification, while soil moisture and temperature sensors provide contextual environmental data to support diagnosis. The robot’s mobility, powered by solar energy, allows for continuous field monitoring with minimal human intervention. Experimental results demonstrate the system’s high classification performance, achieving 99.39% training accuracy, 97.47% validation accuracy, and 97.13% testing accuracy. Furthermore, the model achieved 99.63% overall accuracy, with a Precision of 99.40%, a Recall/Sensitivity of 99.56%, an F1-score of 99.46%, and a Specificity of 99.99% across multiple disease classes. These results highlight the robustness of the proposed approach in real-world agricultural conditions, enabling reliable disease detection and monitoring. The integration of cloud-based monitoring enables farmers to receive real-time alerts and insights, supporting timely and informed decision-making. This cost-effective, scalable, and environmentally sustainable solution has the potential to transform precision agriculture by enhancing early disease detection, reducing pesticide overuse, and improving crop yield and health.

Why it matches plant phenotyping methods植物病害状態を画像と深層学習で検出するロボット型フェノタイピング基盤が研究の中心であり、技術性能も評価している。

abstractThis paper presents a novel, solar-powered autonomous robotic system designed to detect plant diseases in real time using deep learning and IoT technologies.

Plant growth point localization via epoch-based prior annealing. →

Accurate localization of plant growth points is essential for precision agriculture applications, including electro-weeding and laser weeding. While crop and weed detection has been extensively studied, existing methods focus primarily on object-level recognition and often neglect fine-grained growth point localization. To address this limitation, we propose a novel training strategy, epoch-based prior annealing (EPA), which incorporates the excess green minus excess red (ExG-ExR) index as prior knowledge and introduces schedule factor and gain factor to effectively steer keypoint regression. The experimental results show that incorporating EPA improves keypoint localization performance, with mAP50 increasing by 0.024 and mAP50:95 by 0.011, while maintaining bounding box detection performance. The parameter sensitivity experiments confirmed that both excessively strong and weak guidance can hinder training. Furthermore, analysis of parameters and computational cost shows that the additional overhead introduced by the EPA strategy accounts for less than 0.5% of the total, and be considered negligible. In summary, the proposed EPA strategy significantly improves the accuracy, robustness, and generalizability of plant and growth point detection models, offering a practical and scalable solution for precision agricultural applications.

Why it matches plant phenotyping methods植物の生長点を画像から局在化する学習戦略を開発し、その性能を実験的に検証しており、植物器官の表現型取得が中心です。

abstractwe propose a novel training strategy, epoch-based prior annealing (EPA), which incorporates the excess green minus excess red (ExG-ExR) index as prior knowledge and introduces schedule factor and gain factor to effectively steer keypoint regression.

Light Sources in Hyperspectral Imaging Simultaneously Influence Object Detection Performance and Vase Life of Cut Roses. →

Hyperspectral imaging (HSI) is a noncontact camera-based technique that enables deep learning models to learn various plant conditions by detecting light reflectance under illumination. In this study, we investigated the effects of four light sources-halogen (HAL), incandescent (INC), fluorescent (FLU), and light-emitting diodes (LED)-on the quality of spectral images and the vase life (VL) of cut roses, which are vulnerable to abiotic stresses. Cut roses 'All For Love' and 'White Beauty' were used to compare cultivar-specific visible reflectance characteristics associated with contrasting petal pigmentation. HSI was performed at four time points, yielding 640 images per light source from 40 cut roses. The results revealed that the light source strongly affected both the image quality (mAP@0.5 60-80%) and VL (0-3 d) of cut roses. The HAL lamp produced high-quality spectral images across wavelengths (WL) ranging from 480 to 900 nm and yielded the highest object detection performance (ODP), reaching mAP@0.5 of 85% in 'All For Love' and 83% in 'White Beauty' with the YOLOv11x models. However, it increased petal temperature by 2.7-3 °C, thereby stimulating leaf transpiration and consequently shortening the VL of the flowers by 1-2.5 d. In contrast, INC produced unclear images with low spectral signals throughout the WL and consequently resulted in lower ODP, with mAP@0.5 of 74% and 69% in 'All For Love' and 'White Beauty', respectively. The INC only slightly increased petal temperature (1.2-1.3 °C) and shortened the VL by 1 d in the both cultivars. Although FLU and LED had only minor effects on petal temperature and VL, these illuminations generated transient spectral peaks in the WL range of 480-620 nm, resulting in decreased ODP (mAP@0.5 60-75%). Our results revealed that HAL provided reliable, high-quality spectral image data and high object detection accuracy, but simultaneously had negative effects on flower quality. Our findings suggest an alternative two-phase approach for illumination applications that uses HAL during the initial exploration of spectra corresponding to specific symptoms of interest, followed by LED for routine plant monitoring. Optimizing illumination in HSI will improve the accuracy of deep learning-based prediction and thereby contribute to the development of an automated quality sorting system that is urgently required in the cut flower industry.

Why it matches plant phenotyping methods切り花の状態評価に用いるHSIについて、照明条件が画像品質と検出精度に及ぼす影響を比較・検証し、実運用向けの照明戦略を提案しているため、植物フェノタイピング手法が中心である。

abstractwe investigated the effects of four light sources-halogen (HAL), incandescent (INC), fluorescent (FLU), and light-emitting diodes (LED)-on the quality of spectral images and the vase life (VL) of cut roses

Research on Drought Stress Detection in the Seedling Stage of Yunnan Large-Leaf Tea Plants Based on Biomimetic Vision and Chlorophyll Fluorescence Imaging Technology. →

To address the issue of drought level confusion in the detection of drought stress during the seedling stage of the Yunnan large-leaf tea variety using the traditional YOLOv13 network, this study proposes an improved version of the network, MC-YOLOv13-L, based on animal vision. With the compound eye's parallel sampling mechanism at its core, Compound-Eye Apposition Concatenation optimization is applied in both the training and inference stages. Simulating the environmental information acquisition and integration mechanism of primates' "multi-scale parallelism-global modulation-long-range integration," multi-scale linear attention is used to optimize the network. Simulating the retinal wide-field lateral inhibition and cortical selective convergence mechanisms, CMUNeXt is used to optimize the network's backbone. To further improve the localization accuracy of drought stress detection and accelerate model convergence, a dynamic attention process simulating peripheral search, saccadic focus, and central fovea refinement in primates is used. Inner-IoU is applied for targeted improvement of the loss function. The testing results from the drought stress dataset (324 original images, 4212 images after data augmentation) indicate that, in the training set, the Box Loss, Cls Loss, and DFL Loss of the MC-YOLOv13-L network decreased by 5.08%, 3.13%, and 4.85%, respectively, compared to the YOLOv13 network. In the validation set, these losses decreased by 2.82%, 7.32%, and 3.51%, respectively. On the whole, the improved MC-YOLOv13-L improves the accuracy, recall rate and mAP@50 by 4.64%, 6.93% and 4.2%, respectively, on the basis of only sacrificing 0.63 FPS. External validation results from the Laobanzhang base in Xishuangbanna, Yunnan Province, indicate that the MC-YOLOv13-L network can quickly and accurately capture the drought stress response of tea plants under mild drought conditions. This lays a solid foundation for the intelligence-driven development of the tea production sector and, to some extent, promotes the application of bio-inspired computing in complex ecosystems.

Why it matches plant phenotyping methods茶樹の干ばつストレス状態を画像から検出する改良YOLO手法を開発・検証しており、植物状態の取得・推定が研究の中心である。

abstractthis study proposes an improved version of the network, MC-YOLOv13-L, based on animal vision.

Atlas-Based Spatio-temporal MRI Phenotyping of 3D Fungal Spread in Grapevine Wood →

Abstract In perennial crops, inner wood degradation by pathogens often escapes detection until irreversible damage has occurred. Grapevine trunk disease (GTD) is a well-known example in viticulture that alters plants from within, years before foliar symptoms arise, making early assessment difficult. To overcome this limitation, we present a novel non-destructive 3D + t pipeline for Magnetic Resonance Imaging (MRI) spatial quantification and monitoring of early internal tissue degradation resulting from fungal colonization. This pipeline integrates (i) anatomical alignment and rigid time-series registration of volumetric MRI scans, (ii) a generalized cylindrical coordinate transformation for cross-sectional trunk anatomy normalization, (iii) supervised classification to segment water-depleted (diseased/non-functional) regions, and (iv) population-level statistical analyses including construction of population mean images, probabilistic atlases of lesions, and 3D lesion descriptors. Applied to multiple Vitis vinifera cultivars inoculated with a fungal trunk pathogen, our approach enables time-lapse comparisons between cultivar and treatment in vivo. The results reveal consistent early degradation signals across individuals and cultivar-dependent lesion differences. By combining high-resolution MRI with advanced image processing and statistical atlas tools, this method provides a new paradigm for 3D plant phenotyping of internal disease progression. This methodological innovation allows non-invasive quantification of disease development and comparative assessment of host responses in woody plants, demonstrating its potential to advance understanding and management of GTDs.

Why it matches plant phenotyping methodsMRI画像と画像処理・統計アトラスを統合し、ブドウ樹内部の病変・組織劣化を3Dで定量化する植物フェノタイピング手法の開発が中心である。

abstractwe present a novel non-destructive 3D + t pipeline for Magnetic Resonance Imaging (MRI) spatial quantification and monitoring of early internal tissue degradation resulting from fungal colonization.

A lightweight and generalizable deep learning framework for early detection of rice leaf diseases in complex field environments. →

Rice leaf diseases pose a significant and escalating threat to global food security. Timely and accurate detection, particularly in the critical early stages characterized by subtle lesions, is paramount for effective disease management. However, existing solutions often struggle with the complexities of real-world field environments (e.g., variable lighting, occlusions, complex backgrounds), computational constraints on edge devices, and limited generalizability across diverse disease types and plant species. To address these challenges, this study proposes a novel lightweight deep learning framework specifically designed for robust rice leaf disease detection. Our key innovations include: (1) A Multi-branch Large-kernel Fusion Depthwise (MLFD) module enhancing multi-scale contextual feature extraction critical for identifying subtle early lesions; (2) A Multi-scale Dilated Transformer Attention (MDTA) module integrating spatial and channel attention mechanisms to improve feature representation under complex conditions; (3) A Lightweight Detection Head (Lo-Head) optimized with grouped and depthwise convolutions, drastically reducing model complexity without sacrificing accuracy. Crucially, extensive experiments demonstrate the framework's superior performance. On a dedicated rice leaf disease dataset, it achieves a mean Average Precision mAP@0.5:0.95 of 62.62%, outperforming state-of-the-art lightweight detectors including YOLOv5n (56.73%), YOLOv8n (57.41%), YOLOv10n (56.14%), and the baseline YOLOv11n (60.85%), while maintaining low computational demands (6.3 GFLOPs, 2.66M parameters). Significantly, rigorous generalization experiments validate the model's exceptional transferability. Evaluated on independent datasets encompassing potato and tomato leaf diseases, the proposed framework consistently surpasses comparable models in mAP@0.5:0.95, demonstrating its robust capability to detect diseases across different plant species. This combination of high accuracy, computational efficiency, and remarkable cross-crop generalizability positions our framework as a highly promising tool for practical deployment on resource-limited edge devices (e.g., drones, field sensors) in smart agriculture systems, enabling proactive disease surveillance and precision control strategies across diverse crops.

Why it matches plant phenotyping methods植物葉の病徴を画像から検出する深層学習手法の開発と、独立データセットによる性能・汎化性検証が中心であり、植物の病害状態を直接推定するフェノタイピング手法に該当する。

abstractthis study proposes a novel lightweight deep learning framework specifically designed for robust rice leaf disease detection.

Petal to the metal: The slow road to automating large-scale phenology labeling for herbarium specimens →

ABSTRACT Herbarium specimens represent critical historical records of plant phenology, yet automating annotation of reproductive structures remains challenging given the diversity of floral morphologies, specimen age and quality, and image quality. Here, we present a machine learning pipeline that uses an ensemble modeling approach to detect flowers on herbarium specimens and deliver these data to the phenology research community. After testing multiple strategies for generating training data, we found in-house expert-curated annotations were essential for producing reliable results. Expert validation found relatively strong accuracy for detecting present floral structures, but still had moderately high false negative rates. Applying the ensemble to our filtered final image dataset of 22 million records resulted in 11.1 million records labeled with flowers present. However, only 2.9 million of these contained complete metadata necessary for downstream phenology research, highlighting the need for full label digitization efforts. Still, this dataset represents a large compilation of historical herbarium-derived phenology records available as a resource for the phenology community. We end by demonstrating how integrating these machine-labeled records into Phenobase, a publicly-available phenology database, expands taxonomic and temporal coverage for large-scale phenological analyses, and discuss remaining challenges and next steps.

Why it matches plant phenotyping methods植物標本画像から花の存在を自動検出し、精度検証と大規模な phenology データセット化を行う機械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適格です。

abstractwe present a machine learning pipeline that uses an ensemble modeling approach to detect flowers on herbarium specimens and deliver these data to the phenology research community.

A Lightweight Edge-Deployable Framework for Intelligent Rice Disease Monitoring Based on Pruning and Distillation. →

Digital agriculture and smart farming require crop health monitoring methods that balance detection accuracy with computational cost. Rice leaf diseases threaten yield, while field images often contain small multi-scale lesions, variable illumination and cluttered backgrounds. This paper investigates SCD-YOLOv11n, a lightweight detector designed with these constraints in mind. The model replaces the YOLOv11n backbone with a StarNet backbone and integrates a C3k2-Star module to enhance fine-grained, multi-scale feature extraction. A Detail-Strengthened Cross-scale Detection (DSCD) head is further introduced to improve localization of small lesions. On this architecture, we design a DepGraph-based mixed group-normalization pruning rule and apply channel-wise feature distillation to recover performance after pruning. Experiments on a public rice leaf disease dataset show that the compressed model requires 1.9 MB of storage, achieves 97.4% mAP@50 and 76.2% mAP@50:95, and attains a measured speed of 184 FPS under the tested settings. These results provide a quantitative reference for designing lightweight object detectors for rice disease monitoring in digital agriculture scenarios.

Why it matches plant phenotyping methodsイネ葉の病斑を画像から検出・局在化する軽量モデルを開発し、精度・圧縮性能・速度を評価しており、植物病害状態の取得手法が研究の中心である。

abstractThis paper investigates SCD-YOLOv11n, a lightweight detector designed with these constraints in mind.

YOLO-SAM AgriScan: A Unified Framework for Ripe Strawberry Detection and Segmentation with Few-Shot and Zero-Shot Learning. →

Traditional segmentation methods are slow and rely on manual annotations, which are labor-intensive. To address these limitations, we propose YOLO-SAM AgriScan, a unified framework that combines the fast object detection capabilities of YOLOv11 with the zero-shot segmentation power of the Segment Anything Model 2 (SAM2). Our approach adopts a hybrid paradigm for on-plant ripe strawberry segmentation, wherein YOLOv11 is fine-tuned using a few-shot learning strategy with minimal annotated samples, and SAM2 performs mask generation without additional supervision. This architecture eliminates the bottleneck of pixel-wise manual annotation and enables the scalable and efficient segmentation of strawberries in both controlled and natural farm environments. Experimental evaluations on two datasets, a custom-collected dataset and a publicly available benchmark, demonstrate strong detection and segmentation performance in both full-data and data-constrained scenarios. The proposed framework achieved a mean Dice score of 0.95 and an IoU of 0.93 on our collected dataset and maintained competitive performance on public data (Dice: 0.95, IoU: 0.92), demonstrating its robustness, generalizability, and practical relevance in real-world agricultural settings. Our results highlight the potential of combining few-shot detection and zero-shot segmentation to accelerate the development of annotation-light, intelligent phenotyping systems.

Why it matches plant phenotyping methodsイチゴ果実の検出・セグメンテーションを行う画像ベース手法を開発し、複数データセットで性能評価している。単なる収穫対象の位置特定ではなく、果実マスク抽出を行うため、植物表現型取得手法が中心である。

abstractOur approach adopts a hybrid paradigm for on-plant ripe strawberry segmentation

Automated 3D Phenotyping of Maize Plants: Stereo Matching Guided by Deep Learning →

Automated three-dimensional plant phenotyping is an essential tool for non-destructive analysis of plant growth and structure. This paper presents a low-cost system based on stereo vision for depth estimation and morphological characterization of maize plants. The system incorporates an automatic detection stage for the object of interest using deep learning techniques to delimit the region of interest (ROI) corresponding to the plant. The Semi-Global Block Matching (SGBM) algorithm is applied to the detected region to compute the disparity map and generate a partial three-dimensional representation of the plant structure. The ROI delimitation restricts the disparity calculation to the plant area, reducing processing of the background and optimizing computational resource use. The deep learning-based detection stage maintains stable foliage identification even under varying lighting conditions and shadowing, ensuring consistent depth data across different experimental conditions. Overall, the proposed system integrates detection and disparity estimation into an efficient processing flow, providing an accessible alternative for automated three-dimensional phenotyping in agricultural environments.

Why it matches plant phenotyping methods植物の3次元形態を取得・特徴づけるステレオビジョンと深度推定システムの開発が中心であり、明確な植物フェノタイピング手法です。

abstractThis paper presents a low-cost system based on stereo vision for depth estimation and morphological characterization of maize plants.

Optimised MobileNet for very lightweight and accurate plant leaf disease detection →

Abstract The development of accurate and efficient plant disease classification systems is vital for addressing the challenges of climate change and the growing global demand for food. This study presents $$\hbox {V}^2$$ PlantNet, a novel lightweight multi-class classification model based on a modified MobileNet architecture, designed to detect plant leaf diseases across a diverse range of crop types. $$\hbox {V}^2$$ PlantNet employs depthwise separable convolutions to significantly reduce model complexity without compromising accuracy. The architecture integrates Batch Normalization (BN) and Rectified Linear Unit (ReLU) activation after each convolutional layer, while a multi-stage design enhances feature extraction and overall performance. Despite its compact size, comprising only 389,286 parameters and requiring just 1.46 MB of memory, $$\hbox {V}^2$$ PlantNet achieved up to 99% training accuracy, with validation and test accuracies of 97% and 98%, respectively. Across most classes, precision, recall, and F1-scores ranged from 0.97 to 1.0, demonstrating consistent and robust generalization across diverse plant species. These architectural innovations enable $$\hbox {V}^2$$ PlantNet to outperform larger models such as ResNet-50 and Inception V3 in terms of computational efficiency, owing to its smaller model size (1.46 MB), reduced parameter count (389,286), and faster inference time (0.676 s), offering a scalable solution for real-time plant disease detection in 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軽量な深層学習手法を開発し、精度・計算量・推論速度を比較評価しており、フェノタイピング手法が中心である。

abstractThis study presents $$\hbox {V}^2$$ PlantNet, a novel lightweight multi-class classification model based on a modified MobileNet architecture, designed to detect plant leaf diseases across a diverse range of crop types.

Deep learning for sorghum yield forecasting using uncrewed aerial systems and lab-derived imagery. →

The AI revolution, advanced Graphics Processing Units (GPUs), and open-source platforms have enabled Machine Learning (ML) and Deep Learning (DL) algorithms to rapidly and accurately extract phenotypic features from imagery. Such advancements have led to phenotypic digitization and made rapid yield forecasting possible. Yield predictions are critical to assess the merit of genotypes to propel cultivar development. This trial followed a three-replicated Randomized Complete Block Design (RCBD) with 36 diverse sorghum genotypes in 2023 at Ashland Bottoms, Kansas. The field images were captured 6 m above using a DJI M300 drone at 90° nadir and 45° oblique angles. This research trained YOLO and Faster R-CNN (Detectron2) models to harness yield attributes from UAS field and lab images. The YOLO models outperformed the Faster R-CNN in detecting sorghum panicles, achieving a mean average precision at 50 % IoU (mAP@0.50) scores of 0.92-0.98, compared to 0.61-0.89 for Faster R-CNN. Panicle detection from field imagery showed a linear correlation of 0.86 with ground truth field panicle counts. Lab imagery analyses measured panicle area, seed counts, and seed area with correlation coefficients of 0.79, 0.94, and 0.25 with respective ground truth observations. Support Vector Regression (SVR), Random Forest Regression (RFR), and Decision Tree Regression (DTR) were used to predict yield with correlation coefficients of 0.74, 0.71, and 0.78, respectively, and SHapley Additive exPlanation (SHAP) analysis revealed panicle seed count as the primary driver of yield prediction. We observed YOLO models are well-suited for extracting yield-predictive features from pertinent images. Such features can then be incorporated into ML regression models to predict yield per se performance with greater accuracy. The GitHub link is provided in the Data availability section.

Why it matches plant phenotyping methodsUAS・実験室画像から穂数、穂面積、種子数・面積などの植物形質を深層学習で抽出し、検出精度を検証して収量予測へ利用する方法が研究の中心である。

abstractThis research trained YOLO and Faster R-CNN (Detectron2) models to harness yield attributes from UAS field and lab images.

CitriBEiTNet: A Hybrid CNN-Transformer Architecture Combining MobileNetV2 with BEiT's Global Attention for Automated Citrus Leaf Disease Diagnosis →

Citrus farming plays an essential role in agriculture; however, diseases like canker, greening, black spot, and melanose significantly reduce yield and fruit quality. Efficient classification of citrus leaf diseases is important for crop health maintenance and optimal crop yield. Traditional methods for leaf disease detection are slow, labor-intensive, and often inaccurate, which highlights the need for automated solutions. This research presents a novel hybrid approach for identifying citrus diseases by combining a vision transformer with deep learning architectures. Using Bidirectional Encoder Representation from Image Transformers (BEIT) and MobileNetV2 as feature extractors, the proposed model captures distinctive features from images, which are then classified using Support Vector Machine (SVM). The dataset includes four different disease categories and a healthy class. Data augmentation techniques are applied to improve model robustness. The experimental findings demonstrate that CitriBEiTNet achieves a remarkable training accuracy of 99.82% and a testing accuracy of 99.57%, outperforming current leading techniques. This model provides an efficient, scalable, and economical approach for early disease identification, enabling farmers to take preventive measures and improve agricultural yields.

Why it matches plant phenotyping methods柑橘葉画像から病害状態を自動分類する深層学習手法の開発が研究の中心であり、植物の病害表現型を直接推定している。

abstractThis research presents a novel hybrid approach for identifying citrus diseases by combining a vision transformer with deep learning architectures.

ORANGE PLANT LEAF DISEASE DETECTION AND CLASSIFICATION WITH IMAGE PROCESSING USING A DEEP CONVOLUTIONAL NEURAL NETWORK →

The farming of citrus is a crucial component of Pakistan’s fruit-based agricultural economy. But, the foliar diseases citrus canker, black spot, and greening have been posing a constant threat on citrus’s productivity. An optimal solution is an early and accurate detection of these diseases to improve the productivity. Therefore, this paper proposes an automated citrus leaf disease detection and classification framework based on deep convolutional neural networks (DCNNs). The proposed solution has five stages: image acquisition (dataset), preprocessing, data augmentation, deep feature extraction and optimization, and disease classification. Firstly, the images are obtained from a public dataset downloaded from Kaggle. Secondly, preprocessing techniques are used to improve the image quality and shape, thirdly the data augmentation techniques are used to enhance the model generalization, fourthly pre-trained models DenseNet-121, MobileNet, and InceptionV3 with transfer learning technique to extract deep features, and finally Adam optimizer and categorical cross-entropy loss function are used to fine tune the pre-trained models for classifications. The proposed model is evaluated on accuracy, precision, recall, and F1-score metrics. All the models demonstrated robust performance while DenseNet-121 achieved the best performance. The evaluation results assured the robustness of the use of transfer learning-based DCNN in citrus leaf disease detection.

Why it matches plant phenotyping methods柑橘葉の病害状態を画像から直接検出・分類する深層学習ワークフロー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。

abstractTherefore, this paper proposes an automated citrus leaf disease detection and classification framework based on deep convolutional neural networks (DCNNs).

Real-time segmentation and phenotypic analysis of rice seeds using YOLOv11-LA and RiceLCNN. →

Introduction The real-time, accurate detection and classification of rice seeds are crucial for improving agricultural productivity, ensuring grain quality, and promoting smart agriculture. Although significant progress has been made using deep learning, particularly convolutional neural networks (CNNs) and attention-based models, earlier methods such as threshold segmentation and single-grain classification faced challenges related to computational efficiency and latency, especially in high-density seed agglutination scenarios. This study addresses these limitations by proposing an integrated intelligent analysis model that combines object detection, real-time tracking, precise classification, and high-accuracy phenotypic measurement. Methods The proposed model utilizes the lightweight YOLOv11-LA for real-time grain segmentation, which builds upon the YOLOv11 architecture. YOLOv11-LA incorporates several enhancements over YOLOv11, including separable convolutions, CBAM (Convolutional Block Attention Module) attention mechanisms, and module pruning strategies. These modifications not only improve detection accuracy but also significantly reduce the number of parameters by 63.2% and decrease computational complexity by 51.6%. For classification, the model employs a custom-designed, lightweight RiceLCNN classifier. Additionally, the DeepSORT algorithm is employed for real-time multi-object tracking, and sub-pixel edge detection along with dynamic scale calibration mechanisms are applied for precise phenotypic feature measurement. Results Compared to YOLOv11, the YOLOv11-LA model increases the mAP@0.5:0.95 score by 1.9%, showcasing its superior detection performance while maintaining lower computational overhead. The RiceLCNN classifier achieved classification accuracies of 89.78% on private datasets and 96.32% on public benchmark datasets. The system demonstrated high accuracy in measuring phenotypic features such as seed size and roundness, with measurement errors kept within 0.1 millimeters. The DeepSORT algorithm effectively managed multi-object tracking, reducing duplicate identifications and frame loss in real-time. Discussion Experimental validation confirmed that the YOLOv11-LA model outperforms the original YOLOv11 in terms of both detection speed and accuracy, while also maintaining low computational complexity. The integration of the YOLOv11-LA, RiceLCNN, and DeepSORT algorithms, combined with advanced measurement techniques, underscores the model's potential for industrial applications, particularly in enhancing smart agricultural practices.

Why it matches plant phenotyping methodsイネ種子画像からサイズや真円度を抽出するリアルタイム画像解析手法を開発し、精度・速度・測定誤差を検証しており、表現型取得が研究の中心です。

abstractThe proposed model utilizes the lightweight YOLOv11-LA for real-time grain segmentation

AI-Enabled Intelligent System for Automatic Detection and Classification of Plant Diseases Towards Precision Agriculture →

Technology-driven agriculture, or precision agriculture (PA), is indispensable in the contemporary world due to its advantages and the availability of technological innovations. Particularly, early disease detection in agricultural crops helps the farming community ensure crop health, reduce expenditure, and increase crop yield. Governments have mainly used current systems for agricultural statistics and strategic decision-making, but there is still a critical need for farmers to have access to cost-effective, user-friendly solutions that can be used by them regardless of their educational level. In this study, we used four apple leaf diseases (leaf spot, mosaic, rust and brown spot) from the PlantVillage dataset to develop an Automated Agricultural Crop Disease Identification System (AACDIS), a deep learning framework for identifying and categorizing crop diseases. This framework makes use of deep convolutional neural networks (CNNs) and includes three CNN models created specifically for this application. AACDIS achieves significant performance improvements by combining cascade inception and drawing inspiration from the well-known AlexNet design, making it a potent tool for managing agricultural diseases. AACDIS also has Region of Interest (ROI) awareness, a crucial component that improves the efficiency and precision of illness identification. This feature guarantees that the system can quickly and accurately identify illness-related areas inside images, enabling faster and more accurate disease diagnosis. Experimental findings show a test accuracy of 99.491%, which is better than many state-of-the-art deep learning models. This empirical study reveals the potential benefits of the proposed system for early identification of diseases. This research triggers further investigation to realize full-fledged precision agriculture and smart agriculture.

Why it matches plant phenotyping methods植物葉の病徴領域を画像から検出・分類する深層学習手法の開発が中心であり、植物の病害状態を直接推定するため、方法論文として採用する。

abstractwe used four apple leaf diseases (leaf spot, mosaic, rust and brown spot) from the PlantVillage dataset to develop an Automated Agricultural Crop Disease Identification System (AACDIS), a deep learning framework for identifying and categorizing crop diseases.

Sustainable Plant Disease Management with Real-Time Crop Optimization →

Plant diseases significantly threaten global food security, often leading to severe yield losses and unsustainable reliance on chemical usage and pesticides. This paper presents an integrated, real-time system for sustainable plant disease management using Internet of Things (IoT) sensors, deep learning models, and cloud-edge computing. The proposed framework enables early disease detection and adaptive crop optimization by fusing environmental telemetry with AI-driven image diagnostics. Using the PlantVillage dataset and real-world sensor data, the system achieves 99.1% disease detection accuracy, a 27% reduction in pesticide usage, and a 22% improvement in crop yield, a critical metric in assessing the broader effectiveness of plant disease management strategies compared to leading benchmarks. Field trials confirm its efficacy in enhancing farm productivity while minimizing environmental impact. This work demonstrates a practical, scalable solution for precision agriculture that aligns with the principles of sustainability, resilience, and data-driven decision-making.

Why it matches plant phenotyping methods植物画像から病害状態を推定するAI診断とセンサー統合基盤が研究の中心であり、植物病害フェノタイプの実質的な取得・評価を行っている。

abstractThis paper presents an integrated, real-time system for sustainable plant disease management using Internet of Things (IoT) sensors, deep learning models, and cloud-edge computing.

Detection techniques for tomato diseases under non-stationary climatic conditions. →

Tomato growth is highly susceptible to diseases, making accurate identification crucial for timely intervention. While deep learning models like the YOLO family have demonstrated success in detecting diseases in agricultural settings, they typically assume that training and testing data are independently and identically distributed (i.i.d.), which often doesn't hold in real-world scenarios. When pre-trained models are applied to new environments, performance can degrade due to domain shifts. To address this, we propose CTTA-DisDet, a continuous test-time domain adaptation framework for tomato disease detection that adapts models to evolving environments during testing, improving generalization in unseen domains. CTTA-DisDet utilizes a teacher-student architecture where both models share the same structure. Dynamic data augmentation is introduced, involving explicit and implicit augmentations. Explicit augmentation corrupts input images, while implicit augmentation uses large language models (LLMs) to generate new domain data. The teacher model learns generalized knowledge, and the student model mimics the teacher to distill domain-specific information. During testing, pseudo-labels generated by the teacher update the student model. To prevent catastrophic forgetting, a subset of neurons is randomly restored to their original weights during each test-time iteration. The teacher model is continuously updated via exponential moving average (EMA). Experimental results demonstrate that CTTA-DisDet achieves an impressive 67.9% performance in continuously changing cross-domain environments, significantly benefiting practical applications in non-stationary settings.

Why it matches plant phenotyping methodsトマトの病徴を画像から検出するためのドメイン適応手法を提案し、異なる環境で性能評価しており、植物病害状態の表現型取得が中心である。

abstractwe propose CTTA-DisDet, a continuous test-time domain adaptation framework for tomato disease detection that adapts models to evolving environments during testing, improving generalization in unseen domains.

LBS-YOLO: a lightweight model for strawberry ripeness detection. →

Introduction The traditional strawberry picking operation has long relied on manual work. With the aging trend of the population becoming more and more obvious, the application of intelligent picking technology has become an irreversible trend. However, existing recognition methods still face bottlenecks such as suboptimal recognition accuracy and low computational efficiency. To address these issues, this study constructs a lightweight detection model, LBS-YOLO, based on an improved YOLOv11n architecture, significantly the model's accuracy and interference robustness while greatly compressing the parameter quantity. Methods The LBS-YOLO model is built upon YOLOv11n as the baseline network. In order to enhance the ability of backbone network feature representation, the model designs a lightweight LAWDS module. This design combines channel attention with spatial reconstruction operation to optimize the information retention efficiency in the down-sampling process, thus effectively enhancing the multi-scale feature representation ability and gradient flow propagation performance. Then in the feature fusion stage, the model introduces a Bidirectional Feature Pyramid Network (BiFPN), which not only enables cross-scale feature fusion but also achieves adaptive weighting through a learnable weight allocation mechanism. At last, adopts the C3k2_Star module to replace the conventional C3K2 for improved feature representation. Results On the used strawberry dataset, the LBS-YOLO model reached 88.6% mAP@0.5 and 75.8% mAP@0.5:0.95, which were 2.2 and 1.3 percentage points higher than YOLOv11n, respectively. The LBS-YOLO model improves the recall rate from 83.2% of YOLOv11n to 86.4%, and the F1-score from 81.2% to 82.9%. Its computational complexity is 6.6 GFLOPs and its reasoning speed is 260.7 FPS. Even better, LBS-YOLO only needs 3.4MB of storage space and 1.6 million parameters, which are 34.6% and 38% less than YOLOv11n respectively. Discussion The experiment demonstrates that, the LBS-YOLO model can significantly reduce the number of parameters and effectively improve the detection accuracy and operation efficiency. It successfully alleviated the problems of false detection and missed detection, thereby providing reliable technical support for strawberry growth monitoring, maturity identification and automatic picking.

Why it matches plant phenotyping methodsイチゴの成熟度という植物状態を画像から推定する軽量検出モデルを開発・評価しており、フェノタイピング手法が中心である。

abstractthis study constructs a lightweight detection model, LBS-YOLO, based on an improved YOLOv11n architecture

Image-based cotton leaf disease diagnosis using YOLO and faster R-CNN techniques. →

Cotton has, in recent years, become one of the most important cash crops worldwide while being impacted in yield from leaf disease which generally goes unnoticed in the early stage. Detection methods depend on manual efforts producing slow processes and human errors. Automated detection methods establish low accuracies, limited scalability and real time applications. To tackle the research issue, this study proposes the CLD-Net which stands for Cotton Leaf Disease Detection Network a novel deep learning-based framework which combines Faster-RCNN and YOLOv5 algorithms into a single action to achieve ultimately real time detection of accurate diseases the combination helps identify both the high detection speed of YOLOv5 along with Faster-RCNN regional proposal accuracy. The new method is that the compilation of these two modern object detection methods has been compiled and designed specifically for detecting leaf disease across varying environmental conditions. Notable contributions to this method include increases in classification accuracy, processing speed, real time detection making these methods suitable for farmers agronomists and sensor deployment. CLD-Net integrates YOLOv5 and Faster R-CNN, combining real-time detection capability with precise classification, to deliver robust cotton leaf disease identification. Experimental validation on a curated dataset of cotton leaf images demonstrates the superiority of CLD-Net, achieving an accuracy of 96.7%, which surpasses that of traditional models. These results confirm the potential of the proposed approach to revolutionize crop disease detection, leading to timely intervention and increased yield.

Why it matches plant phenotyping methods綿花葉の病害状態を画像から推定する深層学習手法を開発し、画像データセットで性能検証しており、植物フェノタイピング手法が中心です。

abstractthis study proposes the CLD-Net which stands for Cotton Leaf Disease Detection Network a novel deep learning-based framework which combines Faster-RCNN and YOLOv5 algorithms

Deep learning for three-dimensional (3D) plant phenomics →

Plant phenomics, the comprehensive study of plant phenotypes, has gained prominence as a vital tool for understanding the intricate relationships between genotypes and the environment. Image-based plant phenomics has progressed rapidly, and three-dimensional (3D) phenotyping is a valuable extension of traditional 2D phenomics. However, the increased data dimensionality poses challenges to feature extraction and phenotyping. In recent decades, deep learning has led to remarkable progress in revolutionizing 3D phenotyping. Therefore, this review highlights the importance of using deep learning in 3D plant phenomics. It systematically overviews the capabilities of deep learning for 3D computer vision, covering 3D representation, classification, detection and tracking, semantic segmentation, instance segmentation, and generation. Additionally, deep learning techniques for 3D point preprocessing (e.g., annotation, downsampling, and dataset organization) and various plant phenotyping tasks are discussed. Finally, the challenges and perspectives associated with deep learning in 3D plant phenomics are summarized, including (1) benchmark dataset construction by using synthetic datasets and methods such as generative artificial intelligence and unsupervised or weakly supervised learning; (2) accurate and efficient 3D point cloud analysis by leveraging multitask learning, lightweight models, and self-supervised learning; and (3) deep learning for 3D plant phenomics by exploring interpretability, extensibility, and multimodal data utilization. The exploration of deep learning in 3D plant phenomics is poised to spur breakthroughs in a new dimension of plant science.

Why it matches plant phenotyping methods3D植物フェノミクスにおける深層学習手法を体系的にレビューしており、植物形質の抽出・推定手法が中心である。

abstractTherefore, this review highlights the importance of using deep learning in 3D plant phenomics.

Simultaneous Identification on Tomato Variety and Maturity Based on Local and Global Feature Fusion. →

Varieties show their unique characteristics in morphology, growth, and fruits. Tomato maturity is related to multiple dimensional characteristics including color, texture, smell, etc. An effective classification method of tomato variety and maturity is crucial for evaluating its growth and yield. However, due to the complex growth environment, some problems such as leaf occlusion and fruit shaded by each other make it difficult to accurately and efficiently identify them. To solve these problems, this study innovatively proposes a simultaneous detection model on tomato variety and maturity based on improved YOLOv8n, with the combination of frequency-adaptive dilated convolution (FADC) feature extraction module and the high-level screening-feature path aggregation network (HSPAN) with the aim of local and global feature fusion by the channel attention module and feature selection fusion mechanism. In addition, we use the Powerful-IoU (PIoU) loss function to replace the original Complete IoU (CIoU) to enhance the accuracy of bounding boxes. We also introduce a dynamic detection head as the final output of the model, which can adaptively adjust the focus of feature extraction according to the color and size of tomato fruits, thereby improving the recognition accuracy. Experimental results show that our model with better global perception capability achieves the highest detection accuracy and lower computation complexity among the comparative models.

Why it matches plant phenotyping methodsトマト果実の成熟度という観察可能な植物状態を画像から推定する検出モデルを開発しており、特徴抽出・検出ヘッド・損失関数の改良と比較評価が研究の中心である。

abstractthis study innovatively proposes a simultaneous detection model on tomato variety and maturity based on improved YOLOv8n

Multi-Scale Feature Fusion Based RT-DETR for Tomato Leaf Disease Detection in Complex Backgrounds. →

In this study, we propose a multi-scale feature fusion network based on an improved RT-DETR model for the efficient detection of tomato leaf disease. Our model combines the multi-scale extended residual module by capturing contextual information at various scales and the multi-scale feature pyramid network by integrating feature information from different levels, which improves feature extraction capability and reduces the interference of complex backgrounds on feature extraction, thereby improving information transmission efficiency and the accuracy of the model. In addition, the novel loss function called adaptive focal loss (AFL) was used, which is based on traditional focal loss with the introduction of attenuation factors to focus the model's attention to high-loss features to alleviate overfitting and of dynamic weight adjustment mechanisms to focus on the more important features during the training process to improve the overall learning performance. Another significant advantage of AFL is that it can more efficiently improve the detection accuracy on imbalanced datasets than on balanced datasets. These innovations optimized the learning strategy of the model, making AP@0.50 up to 97.9% on detecting the categories of tomato diseases. In addition, this model also achieves the high detection accuracy of 85.4% on other crop diseases. These results provide valuable references for agriculture applications.

Why it matches plant phenotyping methodsトマト葉の病害状態を画像から検出する深層学習手法を提案・改良しており、植物病害表現型の抽出法が研究の中心である。

abstractwe propose a multi-scale feature fusion network based on an improved RT-DETR model for the efficient detection of tomato leaf disease.

GAE-YOLO: a lightweight multimodal detection framework for tomato smart agriculture with edge computing →

Introduction The advancement of smart agriculture has witnessed increasing applications of computer vision in crop monitoring and management. However, existing approaches remain challenged by high computational complexity, limited real-time capability, and poor multi-task coordination in tomato cultivation scenarios. Methods To address these limitations, an intelligent tomato management system is proposed based on the Ghost-based Adaptive Efficient You Only Look Once (GAE-YOLO) algorithm. The lightweight architecture of the GAE-YOLO framework is achieved through the replacement of standard convolutional layers with Ghost Convolution (GhostConv) modules, while detection accuracy is significantly improved by the integration of both AReLU activation functions and Effective Intersection over Union (E-IoU) loss optimization. The system, implemented on a Jetson TX2 embedded platform, also incorporates ZED stereo vision for 3D localization and a PyQt6-based visualization platform. Results When implemented on Jetson TX2, the system achieving 93.5% mean Average Precision at 50% intersection over union (mAP@50) at 10.2 frames per second (FPS), which can be optimized to 27 FPS by employing TensorRT acceleration and 720p resolution for scenarios demanding higher throughput. Furthermore, it establishes standardized assessment systems for tomato maturity and yield prediction, and offers integrated modules for disease diagnosis and agricultural large language model consultation. Discussion This work establishes a new paradigm for edge computing in agriculture while providing critical technical support for smart farming development.

Why it matches plant phenotyping methodsトマトの成熟度・収量予測および病害診断を含む画像・3Dビジョン基盤を開発し、エッジ環境で性能評価しているため、植物表現型取得が中心的な研究である。

abstractan intelligent tomato management system is proposed based on the Ghost-based Adaptive Efficient You Only Look Once (GAE-YOLO) algorithm

TflosYOLO+TFSC: an accurate and robust model for estimating flower count and flowering period. →

Tea flowers play a crucial role in taxonomic research and hybrid breeding of tea plants. As traditional methods of observing tea flower traits are labor-intensive and inaccurate, TflosYOLO and Tea Flowering Stage Classification (TFSC) models were proposed for tea flowering quantification, which enable the estimation of flower count and flowering period. In this study, a highly representative and diverse dataset was constructed by collecting flower images from 29 tea accessions in 2 years. Based on this dataset, the TflosYOLO model was built on the YOLOv5 architecture and enhanced with the Squeeze-and-Excitation (SE) network, Adaptive Rectangular Convolution, and Attention Free Transformer, which is the first model to offer a viable solution for detecting and counting tea flowers. The TflosYOLO model achieved a mean Average Precision at 50% IoU (mAP50) of 0.844, outperforming YOLOv5, YOLOv7, and YOLOv8. Furthermore, the TflosYOLO model was tested on 31 datasets encompassing 26 tea accessions and five flowering stages, demonstrating high generalization and robustness. The correlation coefficient (R 2 ) between the predicted and actual flower counts was 0.964. Additionally, the TFSC model-a seven-layer neural network-was designed for the automatic classification of the flowering period. The TFSC model was evaluated for 2 years and achieved an accuracy of 0.738 and 0.899. Using the TflosYOLO+TFSC model, the tea flowering dynamics were monitored, and the changes in flowering stages were tracked across various tea accessions. The framework provides crucial support for tea plant breeding programs and the phenotypic analysis of germplasm resources.

Why it matches plant phenotyping methods茶花画像から花数と開花期を推定するモデルを開発・検証しており、植物表現型の取得・抽出手法が研究の中心である。

abstractTflosYOLO and Tea Flowering Stage Classification (TFSC) models were proposed for tea flowering quantification, which enable the estimation of flower count and flowering period.

DP-MaizeTrack: a software for tracking the number of maize plants and leaves information from UAV image. →

In modern agricultural production, accurate monitoring of maize growth and leaf counting is crucial for precision management and crop breeding optimization. Current UAV-based methods for detecting maize seedlings and leaves often face challenges in achieving high accuracy due to issues such as low spatial-resolution, complex field environments, variations in plant scale and orientation. To address these challenges, this study develops an integrated detection and visualization software, DP-MaizeTrack, which incorporates the DP-YOLOv8 model based on YOLOv8. The DP-YOLOv8 model integrates three key improvements. The Multi-Scale Feature Enhancement (MSFE) module improves detection accuracy across different scales. The Optimized Spatial Pyramid Pooling-Fast (OSPPF) module enhances feature extraction in diverse field conditions. Experimental results in single-plant detection show that the DP-YOLOv8 model outperforms the baseline YOLOv8 with improvements of 3.9% in Precision (95.1%), 4.1% in Recall (91.5%), and 4.0% in mAP50 (94.9%). The software also demonstrates good accuracy in the visualization results for single-plant and leaf detection tasks. Furthermore, DP-MaizeTrack not only automates the detection process but also integrates agricultural analysis tools, including region segmentation and data statistics, to support precision agricultural management and leaf-age analysis. The source code and models are available at https://github.com/clhclhc/project.

Why it matches plant phenotyping methodsUAV画像からトウモロコシ個体数と葉数を抽出するソフトウェアを開発しており、植物形質取得が研究の中心です。

abstractthis study develops an integrated detection and visualization software, DP-MaizeTrack

WMC-Leafset: A dataset of wax gourd and Mangalore cucumber plants for leaf miner and pest infestation diseased object detection. →

Wax gourd ( Benincasa hispida (Thunb.) Cogn.) and Mangalore Cucumber (Cucumis melo L. subsp. agrestis var. conomon) are nutritionally rich, mineral-dense crops with a short growing cycle, making them a preferred choice for cultivation among farmers across the country. The Mangalore cucumber, also known as the culinary cucumber, Indian yellow cucumber, or Japanese pickling melon, is widely used in Asian cuisine for pickling. While proper nutrient management is essential for optimal growth, disease control poses a significant challenge in ensuring healthy yields, as disease can rapidly spread from one leaf to another, affecting larger areas of the field and reducing crop yield. Since cucurbits grow close to the soil, they spread across the ground, exhibit dense canopies, and often overlap with neighboring plants. Early detection is crucial to ensure sustainable cultivation, food security, and increased crop productivity. To address this challenge, we collected a dataset comprising 3200 images that includes image samples of Wax gourd and Mangalore cucumber plants affected by leaf miner, pests and image samples of healthy leaves. The Cucurbitaceae datasets that are available in the public domain lack representation of the Mangalore cucumber and Wax gourd varieties. To the best of our knowledge, no publicly available dataset exists for the Wax gourd. Moreover, existing datasets typically contain images captured under controlled greenhouse conditions with plain backgrounds, featuring a single leaf per image. They exhibit low background complexity and limit the scope to detect diseases at the object level, including multiple diseases present on a single leaf or plant. The uniqueness of the proposed dataset lies in addressing this gap by providing field-level images of cucurbits. These images capture variations in soil, overlapped leaves, complex background, varying angles and distances, weeds, and human interference. This makes the dataset suitable for training object detection models capable of identifying single and multiple disease instances, and it can also be effectively used for classification tasks to distinguish between healthy and diseased leaves. It supports advancement in deep learning, feature extraction, segmentation and pattern recognition tasks. Additionally, the dataset serves as a valuable resource for plant pathologists, agronomists and agricultural experts in disease detection, monitoring and management, thereby promoting sustainable agricultural practices. By offering open access, this dataset promotes collaboration within the scientific community to facilitate the development of robust disease detection, identification, and disease control, thus enhancing farming practices and increasing agricultural yields and advancing food security.

Why it matches plant phenotyping methods植物の病害状態を画像から検出・分類する公開データセットが研究の中心であり、植物病害の画像ベース表現型解析に該当する。

abstractwe collected a dataset comprising 3200 images that includes image samples of Wax gourd and Mangalore cucumber plants affected by leaf miner, pests and image samples of healthy leaves.

PlantSAM: An object detection-driven segmentation pipeline for herbarium specimens. →

Premise Deep learning-based classification of herbarium images is hampered by background heterogeneity, which introduces noise and artifacts that can potentially mislead models and degrade their accuracy. Addressing these effects is essential to enhance overall performance. Methods We introduce PlantSAM, an automated segmentation pipeline that integrates YOLOv10 for plant region detection and the Segment Anything Model (SAM2) for segmentation. YOLOv10 generates bounding box prompts to guide SAM2, enhancing segmentation accuracy. Both models were fine-tuned on herbarium images and evaluated using intersection over union (IoU) and Sørensen-Dice coefficient metrics. Results PlantSAM achieved state-of-the-art segmentation performance, with an IoU of 0.94 and a Sørensen-Dice coefficient of 0.97. Incorporating segmented images into classification models led to consistent performance improvements across five tested botanical traits, with accuracy gains of up to 4.36% and F1 score improvements of 4.15%. Conclusions Our findings highlight the importance of background removal in herbarium image analysis, as it significantly enhances classification performance by enabling models to focus more effectively on the foreground plant structures.

Why it matches plant phenotyping methods植物画像から背景を除去して植物領域を抽出するセグメンテーション手法の開発・評価が中心であり、植物形質分類への有用性も検証している。

abstractWe introduce PlantSAM, an automated segmentation pipeline that integrates YOLOv10 for plant region detection and the Segment Anything Model (SAM2) for segmentation.

DE-YOLOv13-S: Research on a Biomimetic Vision-Based Model for Yield Detection of Yunnan Large-Leaf Tea Trees. →

To address the challenges of variable target scale, complex background, blurred image, and serious occlusion in the yield detection of Yunnan large-leaf tea tree, this study proposes a deep learning network DE-YOLOv13-S that integrates the visual mechanism of primates. DynamicConv was used to optimize the dynamic adjustment process of the effective receptive field and channel the gain of the primate visual system. Efficient Mixed-pooling Channel Attention was introduced to simulate the observation strategy of 'global gain control and selective integration parallel' of the primate visual system. Scale-based Dynamic Loss was used to simulate the foveation mechanism of primates, which significantly improved the positioning accuracy and robustness of Yunnan large-leaf tea tree yield detection. The results show that the Box Loss, Cls Loss, and DFL Loss of the DE-YOLOv13-S network decreased by 18.75%, 3.70%, and 2.54% on the training set, and by 18.48%, 14.29%, and 7.46% on the test set, respectively. Compared with YOLOv13, its parameters and gradients are only increased by 2.06 M, while the computational complexity is reduced by 0.2 G FLOPs, precision, recall, and mAP are increased by 3.78%, 2.04% and 3.35%, respectively. The improved DE-YOLOv13-S network not only provides an efficient and stable yield detection solution for the intelligent management level and high-quality development of tea gardens, but also provides a solid technical support for the deep integration of bionic vision and agricultural remote sensing.

Why it matches plant phenotyping methods茶樹の収量を画像から検出する深層学習モデルを開発・評価しており、植物形質の取得手法が研究の中心である。

abstractthis study proposes a deep learning network DE-YOLOv13-S

CaneFocus-Net: A Sugarcane Leaf Disease Detection Model Based on Adaptive Receptive Field and Multi-Scale Fusion. →

In the context of global agricultural modernization, the early and accurate detection of sugarcane leaf diseases is critical for ensuring stable sugar production. However, existing deep learning models still face significant challenges in complex field environments, such as blurred lesion edges, scale variation, and limited generalization capability. To address these issues, this study constructs an efficient recognition model for sugarcane disease detection, named CaneFocus-Net, specifically designed for precise identification of sugarcane leaf diseases. Based on a single-stage detection architecture, the model introduces a lightweight cross-stage feature fusion module (CP) to optimize feature transfer efficiency. It also designs a module combining a channel-spatial adaptive calibration mechanism with multi-scale pooling aggregation to enhance the backbone network's ability to extract multi-scale lesion features. Furthermore, by expanding the high-resolution shallow feature layer to enhance sensitivity toward small-sized targets and adopting a phased adaptive nonlinear optimization strategy, detection and localization accuracy along with convergence efficiency have been further improved. Test results on public datasets demonstrate that this method significantly enhances recognition performance for fuzzy lesions and multi-scale targets while maintaining high inference speed. Compared to the baseline model, precision, recall, and mean average precision (mAP50 and mAP50-95) improved by 1.9%, 4.6%, 1.5%, and 1.4%, respectively, demonstrating strong generalization capabilities and practical application potential. This provides reliable technical support for intelligent monitoring of sugarcane diseases in the field.

Why it matches plant phenotyping methodsサトウキビ葉の病斑を画像から検出・局在化する深層学習モデルを開発しており、植物の病害状態を直接推定する手法が研究の中心である。

abstractthis study constructs an efficient recognition model for sugarcane disease detection, named CaneFocus-Net

Artificial Intelligence-Based Plant Disease Classification in Low-Light Environments →

The accurate classification of plant diseases is vital for global food security, as diseases can cause major yield losses and threaten sustainable and precision agriculture. The classification of plant diseases in low-light noisy environments is crucial because crops can be continuously monitored even at night. Important visual cues of disease symptoms can be lost due to the degraded quality of images captured under low-illumination, resulting in poor performance of conventional plant disease classifiers. However, researchers have proposed various techniques for classifying plant diseases in daylight, and no studies have been conducted for low-light noisy environments. Therefore, we propose a novel model for classifying plant diseases from low-light noisy images called dilated pixel attention network (DPA-Net). DPA-Net uses a pixel attention mechanism and multi-layer dilated convolution with a high receptive field, which obtains essential features while highlighting the most relevant information under this challenging condition, allowing more accurate classification results. Additionally, we performed fractal dimension estimation on diseased and healthy leaves to analyze the structural irregularities and complexities. For the performance evaluation, experiments were conducted on two public datasets: the PlantVillage and Potato Leaf Disease datasets. In both datasets, the image resolution is 256 × 256 pixels in joint photographic experts group (JPG) format. For the first dataset, DPA-Net achieved an average accuracy of 92.11% and harmonic mean of precision and recall (F1-score) of 89.11%. For the second dataset, it achieved an average accuracy of 88.92% and an F1-score of 88.60%. These results revealed that the proposed method outperforms state-of-the-art methods. On the first dataset, our method achieved an improvement of 2.27% in average accuracy and 2.86% in F1-score compared to the baseline. Similarly, on the second dataset, it attained an improvement of 6.32% in average accuracy and 6.37% in F1-score over the baseline. In addition, we confirm that our method is effective with the real low-illumination dataset self-constructed by capturing images at 0 lux using a smartphone at night. This approach provides farmers with an affordable practical tool for early disease detection, which can support crop protection worldwide.

Why it matches plant phenotyping methods低照度画像から植物病害状態を分類する新規画像解析モデルを開発し、複数データセットで性能評価しており、植物表現型取得・判定手法が中心である。

abstractwe propose a novel model for classifying plant diseases from low-light noisy images called dilated pixel attention network (DPA-Net).

Cross-scale detection and cross-crop generalization verification of tomato diseases in complex agricultural environments. →

In order to overcome the key challenges associated with detecting tomato leaf disease in complex agricultural environments, such as leaf occlusion, variation in lesion size and light interference, this study presents a lightweight detection model called ToMASD. This model integrates multi-scale feature decoupling and an adaptive alignment mechanism. The model innovatively comprises a dual-branch adaptive alignment module (TAAM) that achieves cross-scale lesion semantic alignment via a dynamic feature pyramid, a local context-aware gated unit (Faster-GLUDet) that uses a spatial attention mechanism to suppress background noise interference, and a multi-scale decoupling detection head (MDH) that balances the detection accuracy of small and diffuse lesions. On a dataset containing six types of disease under various weather conditions, ToMASD achieves an average precision of 84.3%,.by a margin of 4.7% to 12.1% over thirteen mainstream models. The computational load is compressed to 7.1 GFLOPs. Through the introduction of a transfer learning paradigm, the pre-trained weights of the tomato disease detection model can be transferred to common bean and potato detection tasks. Through domain adaptation layers and adversarial feature decoupling strategies, the domain shift problem is overcome, achieving an average precision of 92.7% on the target crop test set. False detection rates in foggy and strong light conditions are controlled at 6.3% and 9.8%, respectively. This study achieves dual breakthroughs in terms of both high-precision detection in complex scenarios and the cross-crop generalization ability of lightweight models. It provides a new paradigm for universal agricultural disease monitoring systems that can be deployed at the edge.

Why it matches plant phenotyping methodsトマト葉の病斑・病害状態を画像から検出するモデルを開発し、複数モデル比較、悪条件評価、他作物への汎化検証を行っており、植物病害フェノタイピング手法が中心である。

abstractthis study presents a lightweight detection model called ToMASD.

EBS-YOLO: edge-optimized bidirectional spatial feature augmentation for in-field detection of wheat Fusarium head blight epidemics. →

Fusarium head blight (FHB), caused by the Fusarium species complex, significantly endangers wheat yield and safety. Accurate and timely assessment of FHB epidemic level in the field is crucial for effective disease management. However, the complex environment and indistinct edges of diseased areas present substantial challenges in distinguishing between healthy and diseased ears, thereby impacting the accuracy of FHB epidemic level detection. This study proposes EBS-YOLO, a novel Edge-Optimized Bidirectional Spatial Feature Augmentation YOLO Network, specifically designed for the rapid and precise determination of FHB epidemic levels at the canopy level. The Focal-Edge Selection Module (FSM) within the backbone replaces original C2f module to enhance edge feature representation and facilitate multi-scale feature extraction. Furthermore, the Dual Spatial-Connection Feature Pyramid Network (DSCFPN), integrating Global-to-Local Spatial Aggregation (GLSA) with bidirectional pyramid interaction, balances global and local feature acquisition while optimizing the feature fusion mechanism. This design enables the model to effectively handle occlusions, scale variations, and complex environments. Experimental results demonstrate substantial improvements over eight comparative models in detecting healthy and diseased wheat ears, achieving mean Average Precision (mAP) of 86.1% and 82.9%, respectively. Notably, the model achieved a mean accuracy of 94.7% in detecting FHB epidemic levels through rigorous spatiotemporal validation using datasets collected from independent fields across different years, underscoring its robust generalization capability. Characterized by its low complexity and lightweight design, EBS-YOLO features a parameter count of 2.05 M, 7.4 GFLOPs, and a model size of 5.0 MB, making it an efficient approach for real-time FHB epidemic level detection.

Why it matches plant phenotyping methods小麦穂の健全・罹病状態と赤かび病の流行レベルを圃場画像から推定する深層学習手法を開発し、独立圃場・異なる年のデータで検証しているため、植物病害フェノタイピング手法が中心である。

abstractThis study proposes EBS-YOLO, a novel Edge-Optimized Bidirectional Spatial Feature Augmentation YOLO Network, specifically designed for the rapid and precise determination of FHB epidemic levels at the canopy level.

An RGB-D Image Dataset for Lychee Detection and Maturity Classification for Robotic Harvesting →

Lychee is a high-value subtropical fruit. The adoption of vision-based harvesting robots can significantly improve productivity while reduce reliance on labor. High-quality data are essential for developing such harvesting robots. However, there are currently no consistently and comprehensively annotated open-source lychee datasets featuring fruits in natural growing environments. To address this, we constructed a dataset to facilitate lychee detection and maturity classification. Color (RGB) images were acquired under diverse weather conditions, and at different times of the day, across multiple lychee varieties, such as Nuomici, Feizixiao, Heiye, and Huaizhi. The dataset encompasses three different ripeness stages and contains 11,414 images, consisting of 878 raw RGB images, 8,780 augmented RGB images, and 1,756 depth images. The images are annotated with 9,658 pairs of lables for lychee detection and maturity classification. To improve annotation consistency, three individuals independently labeled the data, and their results were then aggregated and verified by a fourth reviewer. Detailed statistical analyses were done to examine the dataset. Finally, we performed experiments using three representative deep learning models to evaluate the dataset. It is publicly available for academic

Why it matches plant phenotyping methodsライチ果実の成熟段階という植物器官の状態をRGB-D画像から分類するデータセットを構築し、アノテーション検証と深層学習モデル評価を行っており、表現型取得・評価手法が中心である。

abstractwe constructed a dataset to facilitate lychee detection and maturity classification.

Agri-DSSA: A Dual Self-Supervised Attention Framework for Multisource Crop Health Analysis Using Hyperspectral and Image-Based Benchmarks →

Recent advances in hyperspectral imaging (HSI) and multimodal deep learning have opened new opportunities for crop health analysis; however, most existing models remain limited by dataset scope, lack of interpretability, and weak cross-domain generalization. To overcome these limitations, this study introduces Agri-DSSA, a novel Dual Self-Supervised Attention (DSSA) framework that simultaneously models spectral and spatial dependencies through two complementary self-attention branches. The proposed architecture enables robust and interpretable feature learning across heterogeneous data sources, facilitating the estimation of spectral proxies of chlorophyll content, plant vigor, and disease stress indicators rather than direct physiological measurements. Experiments were performed on seven publicly available benchmark datasets encompassing diverse spectral and visual domains: three hyperspectral datasets (Indian Pines with 16 classes and 10,366 labeled samples; Pavia University with 9 classes and 42,776 samples; and Kennedy Space Center with 13 classes and 5211 samples), two plant disease datasets (PlantVillage with 54,000 labeled leaf images covering 38 diseases across 14 crop species, and the New Plant Diseases dataset with over 30,000 field images captured under natural conditions), and two chlorophyll content datasets (the Global Leaf Chlorophyll Content Dataset (GLCC), derived from MERIS and OLCI satellite data between 2003–2020, and the Leaf Chlorophyll Content Dataset for Crops, which includes paired spectrophotometric and multispectral measurements collected from multiple crop species). To ensure statistical rigor and spatial independence, a block-based spatial cross-validation scheme was employed across five independent runs with fixed random seeds. Model performance was evaluated using R2, RMSE, F1-score, AUC-ROC, and AUC-PR, each reported as mean ± standard deviation with 95% confidence intervals. Results show that Agri-DSSA consistently outperforms baseline models (PLSR, RF, 3D-CNN, and HybridSN), achieving up to R2=0.86 for chlorophyll content estimation and F1-scores above 0.95 for plant disease detection. The attention distributions highlight physiologically meaningful spectral regions (550–710 nm) associated with chlorophyll absorption, confirming the interpretability of the model’s learned representations. This study serves as a methodological foundation for UAV-based and field-deployable crop monitoring systems. By unifying hyperspectral, chlorophyll, and visual disease datasets, Agri-DSSA provides an interpretable and generalizable framework for proxy-based vegetation stress estimation. Future work will extend the model to real UAV campaigns and in-field spectrophotometric validation to achieve full agronomic reliability.

Why it matches plant phenotyping methods植物のクロロフィル含量・活力・病害ストレスを画像／ハイパースペクトルから推定する新規深層学習フレームワークを開発・評価しており、植物表現型の取得・推定が中心である。

abstractthis study introduces Agri-DSSA, a novel Dual Self-Supervised Attention (DSSA) framework that simultaneously models spectral and spatial dependencies through two complementary self-attention branches.

New-generation rice seed germination assessment: high efficiency and flexibility via SeedRuler web-based platform. →

Introduction The germination rate of rice seed is a critical indicator in agricultural research and production, directly influencing crop yield and quality. Traditional assessment methods based on manual visual inspection are often time-consuming, labor-intensive, and prone to subjectivity. Existing automated approaches, while helpful, typically suffer from limitations such as rigid germination standards, strict imaging requirements, and difficulties in handling the small size, dense arrangement, and variable radicle lengths of rice seeds. Methods To address these challenges, we present SeedRuler, a versatile, web-based application designed to improve the accuracy, efficiency, and usability of rice seed germination analysis. SeedRuler integrates three core components: SeedRuler-IP, a traditional image processing-based module; SeedRuler-YOLO, a deep learning model built on YOLOv5 for high-precision object detection; and SeedRuler-SAM, which leverages the Segment Anything Model (SAM) for fine-grained seed segmentation. A dataset of 1,200 rice seed images was collected and manually annotated to train and evaluate the system. An interactive module enables users to flexibly define germination standards based on specific experimental needs. Results SeedRuler-YOLO achieved a mean average precision (mAP) of 0.955 and a mean absolute error (MAE) of 0.110, demonstrating strong detection accuracy. Both SeedRuler-IP and SeedRuler-SAM support interactive germination standard customization, enhancing adaptability across diverse use cases. In addition, SeedRuler incorporates an automated seed size measurement function developed in our prior work, enabling efficient extraction of seed length and width from each image. The entire analysis pipeline is optimized for speed, delivering germination results in under 30 seconds per image. Conclusions SeedRuler overcomes key limitations of existing methods by combining classical image processing with advanced deep learning models, offering accurate, scalable, and user-friendly germination analysis. Its flexible standard-setting and automated measurement features further enhance usability for both researchers and agricultural practitioners. SeedRuler represents a significant advancement in rice seed phenotyping, supporting more informed decision-making in seed selection, breeding, and crop management.

Why it matches plant phenotyping methodsイネ種子の発芽状態と種子サイズを画像から抽出するウェブ型フェノタイピング手法を開発し、データセットで性能評価しているため、方法が研究の中心である。

abstractwe present SeedRuler, a versatile, web-based application designed to improve the accuracy, efficiency, and usability of rice seed germination analysis.

Hybrid deep learning for smart paddy disease diagnosis using self supervised hierarchical reconstruction and attention based temporal analysis →

Accurate and early disease detection in paddy crops is essential for maximizing crop yield which ensures food security. Traditional methods are often labor-intensive, time-consuming, and domain-specific expertise. Feed-forward deep-learning models will perform accurate disease detection through the identification of spatial patterns. However, they cannot predict the diseases at the early stages due to the lack of temporal information. Temporal observations will help perform continuous monitoring and detect minute changes in the crops at the early times. To tackle this problem, we proposed Self-Supervised Deep Hierarchical Reconstruction (SSDHR), and Long Short-Term Memory (LSTM) which perform early disease detection based on the spatial and temporal data respectively. The SSDHR network uses multi-branch convolution kernels to extract distinct discriminative characteristics rather than conventional leaf-based indicators. It incorporates spatial, and temporal-based attention mechanism Symmetric Fusion Attention (SFA) to improve feature selection and XGBoost (XGB) classifier for better stability. According to experimental findings, the suggested framework achieves a 99.25% accuracy rate in identifying and classifying 13 paddy classes, including normal, blast, hispa, tungro, white stem borer, brown spot, leaf roller, downy mildew, yellow stem borer, bacterial leaf blight, bacterial leaf streak, black stem borer, and bacterial panicle blight.

Why it matches plant phenotyping methodsイネ病害の症状を空間・時間画像データから検出・分類する深層学習フレームワークを提案し、その性能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す。

abstractwe proposed Self-Supervised Deep Hierarchical Reconstruction (SSDHR), and Long Short-Term Memory (LSTM) which perform early disease detection based on the spatial and temporal data respectively.

Densenet 169-Based Plant Disease Detection of PlantVillage Dataset →

Abstract Using the PlantVillage dataset and a DenseNet-169-based spatial attention module, this article introducesa unique method for plant disease diagnosis. Because plant diseases can have a major influence onagricultural productivity, prompt intervention depends on precise identification.The intricacies and variances found in plant disease photos are frequently too much for conventionaltechniques to handle. We use DenseNet-169's strong feature extraction capabilities and supplement themwith a spatial attention module that highlights the most pertinent areas of the images in order to overcomethese difficulties. Additionally, we use fine-tuning methods to maximize our model's performance. Byfine-tuning, the DenseNet-169 architecture may better adjust to the unique features of the PlantVillagedataset, increasing its accuracy and resilience. When paired with spatial attention and fine-tuning,DenseNet-169 performs better than baseline models, attaining higher classification accuracy across arange of plant illnesses. Our results demonstrate how well DenseNet-169 may be integrated withfine-tuning and spatial attention processes for the diagnosis of plant diseases. This approach has thepotential to significantly improve crop management and yield by increasing detection accuracy andfostering more automated and dependable agricultural operations.

Why it matches plant phenotyping methods植物病害画像から病害状態を推定する深層学習手法の開発が中心であり、植物表現型（病害状態）の画像ベース推定に該当する。

abstractthis article introducesa unique method for plant disease diagnosis

A SCG-YOLOv8n potato counting framework with efficient mobile deployment. →

Accurately detecting and counting potatoes during early harvest is essential for estimating yield, automating sorting, and supporting data-driven agricultural decisions. However, field environments often present practical challenges-such as soil occlusion, overlapping tubers, and inconsistent lighting-that hinder robust visual recognition. In response, we introduce SCG-YOLOv8n, a compact and field-adapted detection framework built upon the YOLOv8n architecture and specifically tailored for small-object detection in real-world farming conditions. The model incorporates three practical enhancements: a C-SPD module that preserves spatial detail to improve recognition of partially buried tubers; an S-CARAFE operator that reconstructs fine-scale features during upsampling; and GhostShuffleConv layers that reduce computational overhead without sacrificing accuracy. Through extensive field-based experiments, SCG-YOLOv8n consistently outperforms YOLOv5n and its base version across all key metrics. Float16 quantization compresses the model to 3.2 MB, enabling real-time inference on Android devices. We also developed PotatoDetector, a mobile application that demonstrates stable performance in field trials, achieving an RMSE of 1.38 and [Formula: see text] of 0.96 in counting tasks. These results suggest that SCG-YOLOv8n offers a practical and scalable tool for precision agriculture, with potential applicability to other root and tuber crop monitoring scenarios.

Why it matches plant phenotyping methodsジャガイモ塊茎の画像検出・計数を行うモデルとモバイル実装を開発し、圃場で性能検証している。塊茎数という植物器官形質の取得が中心であり、単なる収量測定ではない。

abstractwe introduce SCG-YOLOv8n, a compact and field-adapted detection framework built upon the YOLOv8n architecture and specifically tailored for small-object detection in real-world farming conditions.

FEWheat-YOLO: A Lightweight Improved Algorithm for Wheat Spike Detection. →

Accurate detection and counting of wheat spikes are crucial for yield estimation and variety selection in precision agriculture. However, challenges such as complex field environments, morphological variations, and small target sizes hinder the performance of existing models in real-world applications. This study proposes FEWheat-YOLO, a lightweight and efficient detection framework optimized for deployment on agricultural edge devices. The architecture integrates four key modules: (1) FEMANet, a mixed aggregation feature enhancement network with Efficient Multi-scale Attention (EMA) for improved small-target representation; (2) BiAFA-FPN, a bidirectional asymmetric feature pyramid network for efficient multi-scale feature fusion; (3) ADown, an adaptive downsampling module that preserves structural details during resolution reduction; and (4) GSCDHead, a grouped shared convolution detection head for reduced parameters and computational cost. Evaluated on a hybrid dataset combining GWHD2021 and a self-collected field dataset, FEWheat-YOLO achieved a COCO-style AP of 51.11%, AP@50 of 89.8%, and AP scores of 18.1%, 50.5%, and 61.2% for small, medium, and large targets, respectively, with an average recall (AR) of 58.1%. In wheat spike counting tasks, the model achieved an R 2 of 0.941, MAE of 3.46, and RMSE of 6.25, demonstrating high counting accuracy and robustness. The proposed model requires only 0.67 M parameters, 5.3 GFLOPs, and 1.6 MB of storage, while achieving an inference speed of 54 FPS. Compared to YOLOv11n, FEWheat-YOLO improved AP@50, AP_s, AP_m, AP_l, and AR by 0.53%, 0.7%, 0.7%, 0.4%, and 0.3%, respectively, while reducing parameters by 74%, computation by 15.9%, and model size by 69.2%. These results indicate that FEWheat-YOLO provides an effective balance between detection accuracy, counting performance, and model efficiency, offering strong potential for real-time agricultural applications on resource-limited platforms.

Why it matches plant phenotyping methodsコムギ穂の検出・計数という植物器官形質の画像ベース推定手法を開発し、データセット上で精度・計算効率・堅牢性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 study proposes FEWheat-YOLO, a lightweight and efficient detection framework optimized for deployment on agricultural edge devices.

ADQ-YOLOv8m: a precise detection model of sugarcane disease in complex environment. →

Introduction Current research on sugarcane disease identification primarily focuses on a limited number of typical diseases, often constrained by specific target groups or conditions. To address this, we propose an enhanced ADQ-YOLOv8m model based on the YOLOv8m framework, enabling precise detection of sugarcane diseases. Methods The detection head is modified to a Dynamic Head to enhance feature representation capabilities. Following the Detect module, we introduce the ATSS dynamic label assignment strategy and the QFocalLoss loss function to address issues such as class imbalance, thereby bolstering the model's feature representation capabilities. Results Experimental results demonstrate that ADQ-YOLOv8m outperforms nine other mainstream object detection models, achieving precision, recall, mAP50, mAP50-95, and F1 scores of 86.90%, 85.40%, 90.00%, 77.40%, and 86.00%, respectively. Discussion Finally, comprehensive evaluation of the ADQ-YOLOv8m model's performance is conducted using visual analysis of image predictions and cross-scenario adaptability testing. The experimental results indicate that the proposed model excels in multi-objective processing and demonstrates strong generalization capabilities, suitable for scenarios involving multiple objectives, multiple categories, and class imbalance. The detection method proposed exhibits excellent detection performance and potential, providing robust support for the development of intelligent sugarcane cultivation and disease control.

Why it matches plant phenotyping methodsサトウキビ病害の画像から病害状態を推定する検出モデルを開発・評価しており、植物の病害表現型取得が中心的な方法論的貢献である。

abstractwe propose an enhanced ADQ-YOLOv8m model based on the YOLOv8m framework, enabling precise detection of sugarcane diseases.

Dataset of Ash gourd plant leaf images for detection and classification →

The Ash Gourd dataset is valuable since it was collected from the diverse regions within the district of Dhaka in Bangladesh. This dataset represents one of the first attempts to document, elicit, and categorize the health conditions of Ash Gourd (Benincasa hispida) plants in Bangladesh based on healthy samples, aphid plurality, downy mildew, leaf curl, and leaf miner-infested categories. Ash Gourd is one of the region's most important vegetables because of its nutritional and economic value; thus, it is essential to know diseases' manifestation in the improvement of agricultural productivity. The Ash Gourd dataset contains 2676 images, structured into the five categories of Healthy, Aphid, Downy Mildew, Leaf Curl, and Leaf Miner. All images in all categories are raw which can be used flexibly according to the needs of analysis and model training. Concretely, the Healthy class consists of 803 images, while the four other classes contain 1,873 images. This structured way of collecting data will, in turn, enable deeper analysis and help construct machine learning models for disease classification, hence providing worthy insights into Ash Gourd plant health.

Why it matches plant phenotyping methodsアッシュゴード葉の画像データセットを構築し、植物の健康状態・病徴カテゴリを分類するための再利用可能なデータ資源を提供しており、植物病害状態の画像ベース表現型解析が中心です。

abstractThe Ash Gourd dataset contains 2676 images, structured into the five categories of Healthy, Aphid, Downy Mildew, Leaf Curl, and Leaf Miner.

BudCAM: An Edge-Computing Camera System for Bud Detection in Muscadine Grapevines →

Bud break is a critical phenological stage in muscadine grapevines, marking the start of the growing season and the increasing need for irrigation management. Real-time bud detection enables irrigation to match muscadine grape phenology, conserving water and enhancing performance. This study presents BudCAM, a low-cost, solar-powered, edge-computing camera system based on Raspberry Pi 5 and integrated with LoRa radio board, developed for real-time bud detection. Nine BudCAMs were deployed at Florida A&M University Center for Viticulture and Samll Fruit Research from mid February to mid March, 2024, monitoring three wine cultivars (A-27, noble, and Floriana) with three replicates each. Muscadine grape canopy images were captured every 20 minutes between 7:00 to 19:00, generating 2656 high-resolution (4656×3456 pixels) bud break images as database for bud detection algorithm development. The dataset was divided into 70% training, 15% validation, and 15% test. YOLOv11 models were trained using two primary strategies: a direct single-stage detector on tiled raw images and a refined two-stage pipeline that first identifies the grapevine cordon. Extensive evaluation of multiple model configurations identified top performers for both the single-stage (mAP@0.5=86.0%) and two-stage (mAP@0.5=85.0%) approaches. Further analysis revealed that preserving image scale via tiling was superior to alternative inference strategies like resizing or slicing. Field evaluations during the 2025 growing season confirmed the system’s effectiveness, with the two-stage model showing greater robustness to environmental noise like lens fog. A time-series filter smooths the raw daily counts to reveal a clear phenological trend for visualization. In its final deployment, the autonomous BudCAM system captures an image, runs inference on-device, and transmits the bud count in under three minutes, demonstrating a complete, field-ready solution for precision vineyard management.

Why it matches plant phenotyping methodsブドウの芽数・芽吹きという植物の表現型を、エッジカメラ、画像データセット、検出アルゴリズム、時系列処理で取得・推定するシステムを開発・評価しており、方法が研究の中心である。

abstractThis study presents BudCAM, a low-cost, solar-powered, edge-computing camera system based on Raspberry Pi 5 and integrated with LoRa radio board, developed for real-time bud detection.

Three-dimensional reconstruction of densely planted rice seedlings based on MultiView images. →

Three-dimensional(3D) seedling reconstruction technology can provide critical technical support for monitoring plant growth, phenotyping high-throughput plants, and conducting precision agriculture. However, multiview image-based reconstruction methods, which rely on image registration and feature matching, are susceptible to issues such as similar textures and viewpoint differences, leading to matching errors and the loss of key structural information. This can result in local deficiencies and reduced accuracy in the reconstructed models. Therefore, to attain improved reconstruction accuracy under low-cost constraints, deep learning-based feature extraction and matching methods are employed in this study, the SuperPoint network is utilized to increase the robustness of the feature point detection and description processes, and the LightGlue algorithm is introduced to improve the accuracy and stability of matching. Additionally, to reduce the impact of shooting and platform jitter on image quality, a dedicated plant 3D reconstruction platform is designed and constructed, and a dataset of densely planted rice seedlings under light stress conditions is collected, comprising three factors (light quality, light quantity, and the photoperiod) ​× ​three levels, totaling nine groups. Experimental results show that the proposed method achieves optimal performance in terms of its point cloud completeness and reprojection error. The phenotypic parameters (e.g., plant height) extracted from the reconstruction data are strongly correlated with the actual measurements (R 2 ​= ​0.989, RMSE ​= ​4.54 ​mm), validating the potential of the proposed method for applications related to simulating plant growth processes, analyzing the effects of environmental factors (e.g., light), and optimizing crop cultivation schemes.

Why it matches plant phenotyping methodsマルチビュー画像によるイネ幼苗の3D再構成プラットフォームとデータセットを開発し、再構成精度および抽出形質を実測値と検証しており、表現型取得手法が研究の中心である。

abstractdeep learning-based feature extraction and matching methods are employed in this study

EDDet: efficient deep-fusion and dynamic optimization for small target detection in eggplant diseases. →

With the development of smart agriculture and the growth of the global population, vegetable production is facing the dual challenges of diversified planting environments and increased concealment of diseases. Eggplant, as an important economic crop, has its disease detection accuracy directly affecting yield and quality. However, traditional detection methods fail to effectively capture small diseased areas. To address this issue, this paper proposes an improved deep learning small target detection model-the Efficient Deep-fusion Detection Model (EDDet), which is specifically optimized for the recognition of small diseased spots in eggplant disease detection. In the detection network, we innovatively designed the Pinwheel Fusion Feature Extractor (PFFE) framework, replacing the standard convolutions of the first two layers with Pinwheel Convolutions (PConv). By using asymmetric padding and parallel convolution kernel design, the receptive field is effectively expanded, the ability to capture underlying features is enhanced, and the detection of small diseased areas in eggplants is more precise. In the feature fusion stage, this paper designs a Cross-layer Attention Module (CAM), including Cross-layer Channel Attention (CCA) and Cross-layer Spatial Attention (CSA), which can efficiently interact and fuse features of different scales without additional sampling, alleviating the information loss caused by semantic gaps. In addition, to solve the instability caused by IoU fluctuations in the bounding box regression process, the model introduces Scale-based Dynamic Loss (SD Loss), which dynamically adjusts the loss weight based on the size of the target. By adaptively adjusting the proportion of IoU-based loss and location constraint loss, more precise localization and stable regression of small diseased areas in eggplants are achieved. Experimental results demonstrate that EDDet achieves a notable improvement in mAP50 (85.4%), outperforming the baseline by 2.8%.Importantly, EDDet also Maintains excellent efficiency with only 2.75 M parameters, 9.1 GFLOPs, and a high inference speed of 288.3 FPS, which is 37.5 FPS higher than the baseline.These results highlight the model's strong potential for real-time deployment in complex agricultural scenarios where both precision and speed are critical.

Why it matches plant phenotyping methodsナス葉の病斑という植物の病害状態を画像から検出・局在化する深層学習手法を開発し、精度と速度を実験的に評価しており、表現型取得・抽出手法が中心である。

abstractthis paper proposes an improved deep learning small target detection model-the Efficient Deep-fusion Detection Model (EDDet), which is specifically optimized for the recognition of small diseased spots in eggplant disease detection.

Enhanced Plant Leaf Disease Detection Using CLAHE-Processed Images and Custom CNN Deep Learning Architecture →

This study aims to develop an efficient and accurate deep learning-based model for the classification of plant leaf diseases using Convolutional Neural Networks (CNN). The objective is to automate disease detection in agricultural crops to assist farmers and agricultural experts in early and reliable diagnosis. The model is trained on the publicly available “Plant Village CLAHE Processed Data” dataset, which includes high-resolution RGB images of healthy and diseased plant leaves. Images are preprocessed through resizing (128×128), normalized, and split into training, validation, and test sets. Data augmentation techniques such as flipping, zooming, and rotation are used to improve generalization. A custom CNN architecture comprising convolutional, pooling, dense, and dropout layers is employed and trained using the Adam optimizer. Exploratory Data Analysis (EDA) ensures data quality and balance. The model achieves impressive results, with 93% test accuracy, 91% precision, 93% recall, and an F1-score of 92%, indicating robust performance in identifying diverse plant diseases. Training accuracy reached 94.64% with a validation accuracy of 92.95%, confirming minimal overfitting. These results validate the model’s reliability for practical use in smart farming solutions, especially in mobile or IoT-based applications for real-time disease monitoring and 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CNN手法の開発・性能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適格。

abstractThis study aims to develop an efficient and accurate deep learning-based model for the classification of plant leaf diseases using Convolutional Neural Networks (CNN).

Deep learning model BiFPN-YOLOv8m for tree counting in mango orchards using satellite remote sensing data​. →

Mango is a fruit of great economic importance in India. India is the top mango-producing nation in the world, accounting for over half of global mango output. In order to determine the production capability of the insured orchards, a complete inventory is carried out in situ every three years. The inventory includes counting number of trees, grouping them into yield categories, and assessing damaged ones. Satellite Remote Sensing proves to be a vital tool for estimating ecological parameters such as population density, tree health, volume, biomass, and carbon sequestration rates. The significance of tree counting extends beyond orchard evaluations, playing a vital role in environmental protection, agricultural planning, and crop yield forecast. unfortunately, conventional tree counting methods often require very expensive feature engineering, which leads to more errors as well as lower overall optimization. In order to overcome these obstacles, deep learning-based methods have been used to count trees, exhibiting cutting-edge results in this crucial activity. This paper introduces a novel approach employing deep learning for Image-Based Mango Tree counting in high-resolution satellite imagery data. The proposed model, named Bi-directional Feature Pyramid Network (BiFPN)-YOLOv8m an improved version of YOLOv8, employs object detection to effectively separate, locate, and count mango trees with in orchards. A dataset of 1700 training and 300 testing images of mango orchards with trees of various ages is used to evaluate the various YOLOv8 variants, YOLOv8n, YOLOv8s, YOLOv8m, YOLOv8l, YOLOv8x, including YOLOv9, YOLOv10, and BiFPN-YOLOv8m, with a focus on computational efficiency, accuracy, and speed. Experimental findings show that, even under difficult circumstances, the proposed method continuously outperforms state-of-the-art techniques.

Why it matches plant phenotyping methods衛星画像からマンゴー樹木を分離・位置推定・計数する深層学習手法を開発・評価しており、植物個体数という観測可能な形態・構造形質の抽出が中心である。

abstractThis paper introduces a novel approach employing deep learning for Image-Based Mango Tree counting in high-resolution satellite imagery data.

YOLOv8-FDA: lightweight wheat ear detection and counting in drone images based on improved YOLOv8. →

Introduction Wheat is a vital global staple crop, where accurate ear detection and counting are essential for yield prediction and field management. However, the complexity of field environments poses significant challenges to achieving lightweight yet high-precision detection. Methods This study proposes YOLOv8-FDA, a lightweight detection and counting method based on YOLOv8. The approach integrates RFAConv for enhanced feature extraction, DySample for efficient multi-scale upsampling, HWD for compressed and accelerated model training, and the SDL loss for improved bounding box regression. Results Experimental results on the GWHD dataset show that YOLOv8-FDA achieves a precision of 86.3%, recall of 77.5%, and mAP@0.5 of 84.9%, outperforming the original YOLOv8n by significant margins. The model size is 2.96MB with a computational cost of 8.3 GFLOPs, and it operates at 19.2 FPS, enabling real-time counting with over 97.5% accuracy using cross-row segmentation. Discussion The proposed YOLOv8-FDA model demonstrates strong detection performance, lightweight characteristics, and efficient real-time capability, indicating its high practicality and suitability for deployment in real-world agricultural applications.

Why it matches plant phenotyping methodsコムギ穂の検出・計数という植物器官形質の画像ベース抽出法をYOLOv8改良モデルとして開発し、データセット上で性能検証しているため、方法が研究の中心である。

abstractThis study proposes YOLOv8-FDA, a lightweight detection and counting method based on YOLOv8.

Research on Plant Leaf Disease Detection Method Based on Improved YOLOv11 →

Abstract has not been obtained from indexed metadata or an accessible article page.

Why it matches plant phenotyping methods植物葉の病害検出法そのものの改良を主題としており、植物の病徴・病害状態を画像から推定する方法開発に該当する。

titleResearch on Plant Leaf Disease Detection Method Based on Improved YOLOv11

Deep learning driven, image-based phenotyping of seed processing efficiency in sainfoin ( Onobrychis viciifolia ). →

Introduction: spp.) is a perennial legume traditionally cultivated as a forage crop and is now emerging as a promising candidate for development as a perennial grain legume. Despite its potential, no research has addressed the breeding of sainfoin varieties with superior grain processing properties. Methods: We conducted a multifactorial experiment to evaluate the depodding and dehulling efficiency of five commercially available sainfoin varieties. Seeds were processed using two different methods (belt thresher and impact dehuller) across five sample sizes. A pre-trained Faster R-CNN (Region-based Convolutional Neural Network) object detection model was fine-tuned to identify intact pods, whole seeds, and split seeds from images of the processed mixtures. These predictions were used to calculate processing efficiency (PE) for each variety. A comprehensive power analysis was performed to determine the minimum sample size of sainfoin pods required to detect differences in PE with high statistical power. Results: We observed strong varietal differences in PE, as well as clear effects of the processing method. Belt threshing produced mixtures with more intact pods, while the impact dehuller generated a higher proportion of split seeds. Increasing sample size led to more intact pods across all varieties and methods, and notably decreased seed proportion in belt-threshed samples. Statistical modeling combined with object detection outputs revealed that a minimum of 2 g of pods is required to reliably detect an absolute proportional difference of 0.25 in PE between two breeding lines with 80% power. Discussion: Our findings demonstrate that sainfoin varieties differ significantly in processing efficiency and that processing outcomes depend strongly on both method and sample size. Integrating deep learning-based phenotyping with robust statistical design enables efficient evaluation of processing traits and provides actionable guidelines for breeding programs. While deep learning models offer powerful, cost-effective tools for plant phenotyping, their outputs must be paired with rigorous statistical design to yield reliable and actionable insights for crop improvement.

Why it matches plant phenotyping methods画像からポッド・種子を検出し、処理効率という植物由来形質を算出する深層学習ベースの表現型解析が研究の中心であるため。

titleDeep learning driven, image-based phenotyping of seed processing efficiency in sainfoin

Spatial inheritance patterns across maize ears are associated with alleles that reduce pollen fitness →

Significance Statement Early studies noting uneven spatial distribution of progeny genotypes after pollination support a hypothesis where differences in pollen tube growth rate can bias inheritance. We used computer vision and statistical analysis to show alleles reducing maize pollen fitness are likely to produce statistically significant increasing, decreasing, or curvilinear spatial patterns from the apex of the inflorescence to the base, suggesting that differential pollen tube growth is not the only mechanism at play. Summary Often, more pollen grains land on recipient flowers than there are ovules to fertilize. Consequently, the haploid male gametophyte engages in post-pollination competition, one way that pollen genotype can influence inheritance. The maize ( Zea mays subsp. mays L.) inflorescence (ear), with its elongated stigma and style structures (silks), has a conspicuous spatial heterogeneity, with longer silks at the base of the ear than those at the apex. To evaluate the hypothesis that alleles with reduced pollen fitness influence the spatial distribution of progeny genotypes along the ear, we developed an updated phenotyping platform that maps mutant Ds-GFP kernel phenotypes on the ear via an implementation of the Faster R-CNN machine vision model (EarVision.v2) and a statistical pipeline that evaluates the relationship between kernel position and transmission ratio (EarScape). In our dataset (1384 ears) representing 58 Ds-GFP alleles, none with Mendelian inheritance (0/48) showed any significant pollen-conditioned spatial trend. In contrast, 50% of alleles with a pollen-specific transmission defect (5/10) exhibited significant spatial effects. An insertion into a gene encoding a putative actin-binding protein, base-to-apex gradient1* ( bag1* ), conditions increased mutant transmission at the ear apex relative to the base. Surprisingly, mutant alleles of two other pollen-expressed genes can generate the opposite pattern, decreased mutant transmission toward the ear apex; and two mutant alleles of the sperm-cell attachment factor, gamete expressed2 ( gex2 ), can produce ears with transmission highest at both base and apex. We conclude that pollen fitness mutants have relatively common but heterogenous effects on the spatial distribution of progeny genotypes.

Why it matches plant phenotyping methodsトウモロコシ穂上のカーネル表現型を画像認識でマッピングし、空間的位置と遺伝子型伝達比を解析するプラットフォームおよび統計パイプラインを開発しており、表現型取得・抽出法が研究の中心である。

abstractwe developed an updated phenotyping platform that maps mutant Ds-GFP kernel phenotypes on the ear via an implementation of the Faster R-CNN machine vision model (EarVision.v2) and a statistical pipeline that evaluates the relationship between kernel position and transmission ratio (EarScape).

AngleCam V2: Predicting leaf inclination angles across taxa from daytime and nighttime photos →

Understanding how plants capture light and maintain their energy balance is crucial for predicting how ecosystems respond to environmental changes. By monitoring leaf inclination angle distributions (LIADs), we can gain insights into plant behaviour that directly influences ecosystem functioning. LIADs affect radiative transfer processes and reflectance signals, which are essential components of satellite-based vegetation monitoring. Despite their importance, scalable methods for continuously observing these dynamics across different plant species throughout day-night cycles are limited. We present AngleCam V2, a deep learning model that estimates LIADs from both RGB and near-infrared (NIR) night-vision imagery. We compiled a dataset of over 4,500 images across 200 globally distributed species to facilitate generalization across taxa. Moreover, we developed a method to simulate pseudo-NIR imagery from RGB imagery to enable an efficient training of a deep learning model for tracking LIADs across day and night. The model is based on a vision transformer architecture with mixed-modality training using the RGB and the synthetic NIR images. AngleCam V2 achieved substantial improvements in generalization compared to AngleCam V1 (R 2 = 0.62 vs 0.12 on the same holdout dataset). Phylogenetic analysis across 100 genera revealed no systematic taxonomic bias in prediction errors. Testing against leaf angle dynamics obtained from multitemporal terrestrial laser scanning demonstrated the reliable tracking of diurnal leaf movements (R 2 = 0.61-0.75) and the successful detection of water limitation-induced changes over a 14-day monitoring period. This method enables continuous monitoring of leaf angle dynamics using conventional cameras, enabling applications in ecosystem monitoring networks, plant stress detection, interpreting satellite vegetation signals, and citizen science platforms for global-scale understanding of plant structural responses.

Why it matches plant phenotyping methods葉の傾斜角分布という植物形質を画像から推定する深層学習手法を開発し、大規模データセット、既存モデル比較、レーザースキャンによる検証、水ストレス下での追跡評価まで実施しており、フェノタイピング手法が研究の中心です。

abstractWe present AngleCam V2, a deep learning model that estimates LIADs from both RGB and near-infrared (NIR) night-vision imagery.

A novel efficient eggplant disease detection method with multi-scale learning and edge feature enhancement. →

In the context of the rapid development of smart agriculture, the detection of crop diseases remains a critical and challenging task. The diversity in eggplant disease scales, disease edge features, and the complexity of planting backgrounds significantly impact disease detection effectiveness. To address these challenges, we propose an eggplant disease detection network with edge feature enhancement based on multi-scale learning. The overall network adopts a "backbone-neck-head" architecture: the backbone extracts features, the neck performs feature fusion, and a three-scale detection head produces the final predictions. First, we designed the Multi-scale Edge Information Enhance (CSP-MSEIE) module to extract features from different disease scales and highlight edge information to obtain richer target features. Second, the Multi-source Interaction Module (MSIM) and Dynamic Interpolation Interaction Module (DIIM) sub-modules were designed further to enhance the model's capacity for multi-scale feature representation. By leveraging dynamic interpolation and feature fusion strategies, these sub-modules significantly improved the model's ability to detect targets in complex backgrounds. Then, leveraging these sub-modules, we designed the Multi-scale Context Reconstruction Pyramid Network (MCRPN) to facilitate spatial feature reconstruction and hierarchical context extraction. This framework efficiently combines feature information across multiple levels, strengthening the model's ability to capture and utilize contextual details. Finally, we validated the effectiveness of the proposed model on two disease datasets. It is noteworthy that on the eggplant disease data, the proposed disease detection model achieved improvements of 4.7% and 7.2% in mAP50 and mAP50-95 metrics, respectively, and the model's frames per second (FPS) reached 270.5. This detection network provides an effective solution for the efficient detection of crop diseases.

Why it matches plant phenotyping methods植物病害の画像ベース検出ネットワークを開発し、病害状態の推定をデータセットで検証しており、表現型取得・抽出手法が中心である。

abstractwe propose an eggplant disease detection network with edge feature enhancement based on multi-scale learning.

Sugarcane stem node detection with algorithm based on improved YOLO11 channel pruning with small target enhancement. →

Sugarcane stem node detection is critical for monitoring sugarcane growth, enabling precision cutting, reducing spuriousness, and improving breeding for resistance to downfall. However, in complex field environments, sugarcane stem nodes often suffer from reduced detection accuracy due to background interference and shadowing effects. For this reason, this paper proposes an improved sugarcane stem node detection model based on YOLO11. This study incorporates the ASF-YOLO (Attentional Scale Sequence Fusion based You Only Look Once) mechanism to enhance the feature fusion layer of YOLO11. Additionally, a high-resolution detection layer, P2, is integrated into the fusion module to improve the model's ability to detect small objects-particularly sugarcane stem nodes-and to better handle multi-scale feature representations. Secondly, to better align with the P2 small-object detection layer, this paper adopts a shared convolutional detection head named LSDECD (Lightweight Shared Detail-Enhanced Convolutional Detection Head), which can better deal with small target detection while reducing the number of model parameters through parameter sharing and detail-enhanced convolution. Using soft-NMS (non-maximum suppression) to replace the original NMS and combining with Shape-IoU, a bounding box regression method that focuses on the shape and scale of the bounding box itself, makes the bounding box regression more accurate, and solves the problem of the impact of detection caused by occlusion and illumination. Finally, to address the increased complexity introduced by the addition of the P2 detection layer and the replacement of the detection head, channel pruning is applied to the model, effectively reducing its overall complexity and parameter count. The experimental results show that the model before pruning has 96.1% and 53.2% mean average precision mAP50 and mAP50:95, respectively, which are 11.9% and 11.1% higher than the original YOLO11n, and the model after pruning also has 10.8% and 9.3% higher than the original YOLO11n, respectively, and the number of parameters is reduced to 279,778, and model size is reduced to 1.3MB. The computational cost decreased from 11.6 GFlops to 6.6 GFlops.

Why it matches plant phenotyping methodsサトウキビ茎節という植物器官の検出を対象に、改良YOLOモデルの開発と性能評価を中心的に行っており、再利用可能な画像ベース表現型取得手法に該当する。

abstractThe experimental results show that the model before pruning has 96.1% and 53.2% mean average precision mAP50 and mAP50:95

Cott-ADNet: Lightweight Real-Time Cotton Boll and Flower Detection Under Field Conditions →

Cotton is one of the most important natural fiber crops worldwide, yet harvesting remains limited by labor-intensive manual picking, low efficiency, and yield losses from missing the optimal harvest window. Accurate recognition of cotton bolls and their maturity is therefore essential for automation, yield estimation, and breeding research. We propose Cott-ADNet, a lightweight real-time detector tailored to cotton boll and flower recognition under complex field conditions. Building on YOLOv11n, Cott-ADNet enhances spatial representation and robustness through improved convolutional designs, while introducing two new modules: a NeLU-enhanced Global Attention Mechanism to better capture weak and low-contrast features, and a Dilated Receptive Field SPPF to expand receptive fields for more effective multi-scale context modeling at low computational cost. We curate a labeled dataset of 4,966 images, and release an external validation set of 1,216 field images to support future research. Experiments show that Cott-ADNet achieves 91.5% Precision, 89.8% Recall, 93.3% mAP50, 71.3% mAP, and 90.6% F1-Score with only 7.5 GFLOPs, maintaining stable performance under multi-scale and rotational variations. These results demonstrate Cott-ADNet as an accurate and efficient solution for in-field deployment, and thus provide a reliable basis for automated cotton harvesting and high-throughput phenotypic analysis. Code and dataset is available at https://github.com/SweefongWong/Cott-ADNet.

Why it matches plant phenotyping methods綿花の花・ボール認識を対象とする画像解析手法を開発し、データセット作成、外部検証、性能評価まで行っており、植物器官の表現型取得が中心である。

abstractWe propose Cott-ADNet, a lightweight real-time detector tailored to cotton boll and flower recognition under complex field conditions.

Integrating Squeeze-and-Excitation Network with Pretrained CNN Models for Accurate Plant Disease Detection →

The increasing global population and the challenges posed by climate change have intensified the demand for sustainable food production. Traditional agricultural practices are often insufficient, leading to significant crop losses due to diseases and pests, despite the widespread use of pesticides and other chemical interventions. This paper introduces a new approach that integrates deep learning techniques, specifically Convolutional Neural Networks (CNNs) with Squeeze and Excitation (SE) networks, to enhance the accuracy of disease detection in fig leaves. By leveraging three pre-trained CNN models—MobileNetV2, InceptionV3, and Xception—this framework addresses data scarcity issues and improves feature representation while minimizing the risk of overfitting. Data augmentation techniques were employed to counteract data imbalance, and visualization tools like Grad-CAM and t-SNE were utilized for model interpretability. The proposed CNN-SE model was trained and evaluated on a fig leaf dataset comprising 1,196 images of healthy and diseased fig leaves, achieving an accuracy of 92.90% with MobileNet-SE, 91.48% with Inception-SE, and 89.62% with Xception-SE. Our model demonstrates superior performance in detecting fig leaf diseases, presenting a robust solution for sustainable agriculture by providing accurate, efficient, and scalable disease management in crops. The code of the proposed framework is available at https://github.com/lafta/SE-block-with-CNN-Models-for-Plant-Disease-Detection.

Why it matches plant phenotyping methodsイチジク葉の画像から健全・罹病状態を推定する深層学習手法を開発・評価しており、植物病害表現型の取得・分類が中心である。

abstractThis paper introduces a new approach that integrates deep learning techniques, specifically Convolutional Neural Networks (CNNs) with Squeeze and Excitation (SE) networks, to enhance the accuracy of disease detection in fig leaves.

YOLO-LF: application of multi-scale information fusion and small target detection in agricultural disease detection. →

With the increasing threat of agricultural diseases to crop production, traditional manual detection methods are inefficient and highly susceptible to environmental factors, making an efficient and automated disease detection method urgently needed. Existing deep learning models still face challenges in detecting small targets and recognizing multi-scale lesions in complex backgrounds, particularly in terms of multi-feature fusion. To address these issues, this paper proposes an improved YOLO-LF model by introducing modules such as CSPPA (Cross-Stage Partial with Pyramid Attention), SEA (SeaFormer Attention), and LGCK (Local Gaussian Convolution Kernel), aiming to improve the accuracy and efficiency of small target disease detection. Specifically, the CSPPA module enhances multi-scale feature fusion, the SEA module strengthens the attention mechanism for contextual and local information to improve detection accuracy, and the LGCK module increases the model's sensitivity to small lesion areas. Experimental results show that the proposed YOLO-LF model achieves significant performance improvements on the Plant Pathology 2020 - FGVC7 and Plant Pathology 2021 - FGVC8 datasets, particularly in mAP@0.5% and mAP@0.5-0.95%, outperforming existing mainstream models. These results indicate that the proposed method effectively handles complex backgrounds and small target detection tasks in agricultural disease detection, demonstrating high practical value.

Why it matches plant phenotyping methods植物病斑という植物の病害状態を画像から検出・推定する深層学習モデルを開発し、複数データセットで性能評価しているため、植物フェノタイピング手法が中心です。

abstractthis paper proposes an improved YOLO-LF model by introducing modules such as CSPPA (Cross-Stage Partial with Pyramid Attention), SEA (SeaFormer Attention), and LGCK (Local Gaussian Convolution Kernel), aiming to improve the accuracy and efficiency of small target disease detection.

Dual-Path Attention Fusion Network with Adaptive Quantum Monarch Butterfly Optimization for Banana Plant Disease Detection →

Diagnosis of banana plant disease is a crucial aspect of sustaining the harvest of crops and their quality. Visual inspection of certain diseases like Black Sigatoka, Panama disease, and aphids is not easy and can lead to misjudgments. Generally, traditional deep learning approaches have been previously used but they have not performed well in addressing issues of class imbalance, sensitive disease differentiation and noisy images obtained in the field. Furthermore, most models are based on a collection of predetermined preprocessing methods and single-path networks that limit their ability to generalize to a wide variety of environments. Current methods of deep learning tend to achieve reasonable overall performance but fail to perform well on key performance indicators such as recall and F1-score when considering underrepresented and overlapping classes, such as Yellow and Black Sigatoka. Such constraints impede efficient field implementation, as diseases of minority classes are often falsely classified. To overcome these deficiencies, we develop a novel Duel-Path Attention Fusion Network (DPAFNet) that is trained utilizing adaptive quantum monarch butterfly optimization (AQMBO). The concept behind the proposed model is to feed MaxViT and HorNet-S two feature extractors to deliver global contextual details and minute-scale textural features. The traditional filters which do a reasonable job in handling dynamic noise and contrast are replaced by a learnable preprocessing unit. The cross-layer fusion attention encourages interclass discriminative learning of diseased plants. The suggested model has been trained and tested on an open-source dataset of Mendeley banana disease, which includes 5,170 images in 7 disease categories and 1 control condition. The accuracy, F1-score and MCC of 98.6% and 0.93 and 0.87 respectively (achieved experimentally) demonstrate the superiority of DPAFNet over baseline models such as EfficientNetB0 (accuracy 95.0%), DenseNet121 and ResNet50 (accuracy 93.50% and 92.0% respectively). As can be seen, the model had a 0.26-0.48 increase in F1-score in the challenging Panama disease category. These results prove that the proposed architecture can be successfully used to achieve high-accuracy disease classification in smart agriculture that is robust and prepared for field implementation.

Why it matches plant phenotyping methodsバナナ植物の病徴画像から病害状態を推定する深層学習手法を開発し、データセットとベースラインで性能検証しており、植物フェノタイピング手法が中心である。

abstractTo overcome these deficiencies, we develop a novel Duel-Path Attention Fusion Network (DPAFNet) that is trained utilizing adaptive quantum monarch butterfly optimization (AQMBO).

Retrospective image analysis for long-term demography using Google Earth imagery →

1. Ecosystems are rapidly degrading. Widely used approaches to monitor ecosystems to manage them effectively are both expensive and time consuming. The recent proliferation of publicly available imagery from satellites, Google Earth, and citizen-science platforms holds the promise to revolutionising ecological monitoring and optimising their efficiency. However, the potential of these platforms to detect species and track their population dynamics remains under-explored. 2. We introduce a fast, inexpensive method for retrospective image analysis combining current ground-truth data with historical RGB imagery from Google Earth to extract long-term demographic data. We apply this method to three case studies involving two major Mediterranean invasive plant taxa with contrasting growth forms. Specifically, we: (1) utilise deep learning to automatically detect individuals of prickly pear ( Opuntia sp.) across various Mediterranean habitats and image resolutions; (2) reconstruct 10 years of spatially explicit recruitment rates for Opuntia along a climatic gradient; and (3) quantify nearly 20 years of growth dynamics for the clonal invader Carpobrotus sp. in two contrasting environments. 3. Our object detection model, trained with Google Earth imagery, achieves 60-80% success in identifying individuals of Opuntia , regardless of habitat type. Model performance increases with target species colour consistency and contrast, as well with the usage of basic data augmentation techniques. Detection is constrained by individual area (<4 m 2 ) but captures 80% of the examined population. 4. Beyond detection, our time-series analysis of publicly available imagery enables detailed population monitoring. With 10-year image series available for Spain, Greece, and the UK, and 20 years for Portugal, we successfully estimate annual recruitment and growth rates and their climatic sensitivity, identify productive and unproductive years, estimate individual age, characterise population structure, model size-age relationships, and identify recruitment hotspots for targeted management. 5. Our pipeline opens new avenues for cost-effective, large-scale demographic monitoring by retrospectively harnessing open-access imagery. While demonstrated here with invasive plants, we discuss the broad applicability of our approach across taxa and ecosystems. The use of retrospective image analysis for long-term demography with Google Earth imagery has the potential to expedite conservation decisions, support effective restoration, and enable robust ecological forecasting in the Anthropocene.

Why it matches plant phenotyping methodsGoogle Earth画像と深層学習を用いて植物個体の検出、成長・加入率・年齢などの形態・動態形質を抽出する再利用可能な解析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。

abstractWe introduce a fast, inexpensive method for retrospective image analysis combining current ground-truth data with historical RGB imagery from Google Earth to extract long-term demographic data.

A labeled image dataset of common tomato diseases for classification and object detection. →

Computer vision has emerged as a critical enabler of sustainable production in protected agriculture by offering efficient and non-invasive crop disease diagnosis. The development of accurate disease recognition models relies heavily on the availability of high-quality image datasets. This study introduces a tomato disease image dataset collected in 2024 from greenhouse facilities within a modern agricultural park in Sichuan Province, China. The dataset comprises 1026 high-resolution images, including 417 images of viral disease, 82 images of gray mold, and 527 images of bacterial wilt, totaling approximately 2.78 GB. Captured under real-world greenhouse conditions and from multiple angles and distances, the images effectively capture multi-scale phenotypic disease features. Manual annotation was conducted using the LabelImg tool under the guidance of plant pathology experts, with labeled regions covering leaves, fruits, and stems. Annotation files are stored in XML format, each corresponding to a specific image. This dataset is well-suited for research in disease classification, object detection, and phenotyping, and supports deep learning model training and cross-crop transfer learning applications.

Why it matches plant phenotyping methodsトマト病害の症状を画像で捉え、分類・検出モデル用に専門家アノテーションした再利用可能なデータセットであり、植物病害状態の表現型取得が中心である。

abstractThe development of accurate disease recognition models relies heavily on the availability of high-quality image datasets.

Robotic 3D Flower Pose Estimation for Small-Scale Urban Farms →

The small scale of urban farms and the commercial availability of low-cost robots (such as the FarmBot) that automate simple tending tasks enable an accessible platform for plant phenotyping. We have used a FarmBot with a custom camera end-effector to estimate strawberry plant flower pose (for robotic pollination) from acquired 3D point cloud models. We describe a novel algorithm that translates individual occupancy grids along orthogonal axes of a point cloud to obtain 2D images corresponding to the six viewpoints. For each image, 2D object detection models for flowers are used to identify 2D bounding boxes which can be converted into the 3D space to extract flower point clouds. Pose estimation is performed by fitting three shapes (superellipsoids, paraboloids and planes) to the flower point clouds and compared with manually labeled ground truth. Our method successfully finds approximately 80% of flowers scanned using our customized FarmBot platform and has a mean flower pose error of 7.7 degrees, which is sufficient for robotic pollination and rivals previous results. All code will be made available at https://github.com/harshmuriki/flowerPose.git.

Why it matches plant phenotyping methodsカスタムカメラ付きロボットによる3D花姿勢推定アルゴリズムとプラットフォームを開発・評価しており、花の姿勢という植物形質の取得が中心である。

abstractenable an accessible platform for plant phenotyping

FHBDSR-Net: automated measurement of diseased spikelet rate of Fusarium Head Blight on wheat spikes. →

) disease that threatens global food security, requires precise quantification of diseased spikelet rate (DSR) as a phenotypic indicator for resistance breeding. Most techniques for measuring DSR rely on manual spikelet-by-spikelet observation and counting, which is inefficient and destructive. Although deep learning offers great promise for automated DSR measurement, existing intelligent detection algorithms are hampered by the lack of spikelet-level annotated data, insufficient feature representation for diseased spikelets, and weak spatial encoding of densely arranged spikelets. To address these challenges, we constructed a dataset of 620 high-resolution RGB images of wheat spikes with 5,222 spikelet-level annotations to systematically analyze spikelet size distributions to fill small-object detection data gaps in this field. We designed FHBDSR-Net, a light framework for automated DSR measurement centered on diseased spikelet detection, which features (1) multi-scale feature enhancement architecture that dynamically combines lesion textures, morphological features, and lesion-awn contrast through adaptive multi-scale kernels to suppress background noise; (2) the Inner-EfficiCIoU loss function to reduce small-target localization errors in dense contexts; and (3) a scale-aware attention module using dilated convolutions and self-attention to encode multi-scale pathological patterns and spatial distributions to enhance dense spikelet resolution. FHBDSR-Net detected diseased spikelets with an average precision of 93.8% with a lightweight design of 7.2 M parameters. The results were strongly correlated with expert evaluations, with a Pearson correlation coefficient of 0.901. Our method is suitable for deployment on resource-constrained mobile devices, facilitating portable plant phenotyping and smart breeding.

Why it matches plant phenotyping methodsコムギ穂の罹病小穂率という植物病害形質を画像から自動推定する手法を開発し、データセット構築と専門家評価による検証を行っており、フェノタイピング手法が中心である。

abstractrequires precise quantification of diseased spikelet rate (DSR) as a phenotypic indicator for resistance breeding.

PhenoRob-F: An autonomous ground-based robot for high-throughput phenotyping of field crops →

Understanding the genetic basis of quantitative traits related to crop growth, yield, and stress response requires the acquisition of large-scale, high-quality phenotypic datasets. High-throughput phenotyping platforms have become effective tools for meeting this requirement. Autonomous mobile robots have gained prominence owing to their ability to carry heavy payloads, their operational flexibility, and their proximity to crops, which allows for higher imaging resolution. In this study, we introduce PhenoRob-F (a phenotyping robot for the field), a cross-row, wheeled robot designed for efficient and automated phenotyping under field conditions. The mobile platform and phenotyping module of the robot were engineered to meet the specific demands of field phenotyping, with integrated visual and satellite navigation systems enabling autonomous operation. We validated the performance of the robot through a series of experiments involving various crop canopies. By capturing RGB images of rice and wheat, we independently performed wheat ear detection and rice panicle segmentation. For wheat ear detection, we achieve a precision of 0.783, a recall of 0.822, and a mean average precision (mAP) of 0.853 when the YOLOv8m model is used. For rice panicle segmentation, the SegFormer_B0 model yielded a mean intersection over union (mIoU) of 0.949 and an accuracy of 0.987. Additionally, by capturing RGB-D data of maize canopies, we performed 3D reconstructions to calculate plant height, achieving an R 2 of 0.99 compared with manual measurements. Similar experiments with rapeseed yielded an R 2 of 0.97. Near-infrared spectral data collected from drought-stressed rice plants enabled the classification of drought severity into five categories, with classification accuracies ranging from 0.977 to 0.996. Our results reveal that PhenoRob-F is an effective tool for high-throughput phenotyping and is capable of providing precise data to support phenotypic trait analysis and the selection of superior crop genotypes.

Why it matches plant phenotyping methods圃場用自律ロボットと複数の画像・分光センシング、形質抽出手法を開発し、作物キャノピーで性能検証しているため、植物フェノタイピング手法が研究の中心である。

abstractwe introduce PhenoRob-F (a phenotyping robot for the field), a cross-row, wheeled robot designed for efficient and automated phenotyping under field conditions.

MFDN: an efficient detection method for Alstroemeria Genus flowers based on multi-scale feature fusion. →

As an ornamental plant, Alstroemeria Genus Morado holds great significance in precision agriculture for the automatic detection and classification of its flower maturity. However, due to its diverse morphologies, complex growth environments, and factors such as occlusion and lighting changes, related tasks face numerous challenges, and research in this area is relatively scarce. This study proposes a deep - learning - based object detection framework, the Morado Flower Detection Network (MFDN), which consists of two parts: a backbone network and a head network. Novel modules such as C3k2_PPA are introduced. Through multi - branch fusion and the attention mechanism, the ability to detect small targets is enhanced. The head network uses the CARAFE module for upsampling, combines features through Concat, accelerates processing with the optimized C2f module, and finally achieves precise detection and classification through the Detect module. In the comparative experiment on the morado_5may dataset, MFDN performs outstandingly in indicators such as Precision, Recall, and F1 - score. The mean Average Precision (mAP) of MFDN is 1.3% - 5.8% higher than that of YOLO - series models. It has strong generalization ability and is expected to contribute to improving the efficiency and automation level of agricultural production.

Why it matches plant phenotyping methodsアルストロメリア花の成熟度を画像から検出・分類する深層学習手法を開発し、データセット上で比較評価しており、植物表現型取得が中心である。

abstractThis study proposes a deep - learning - based object detection framework, the Morado Flower Detection Network (MFDN)

A lightweight and explainable CNN model for empowering plant disease diagnosis →

Abstract Crop disease is a significant challenge in agriculture, requiring quick and precise detection to safeguard yields and reduce economic losses. Traditional diagnostic methods are slow, labor-intensive, and rely on expert knowledge, limiting scalability for large-scale operations. To overcome these challenges, a novel architecture called Mob-Res , combining residual learning with the MobileNetV2 feature extractor, is introduced in this work. Despite having only 3.51 million parameters, Mob-Res is lightweight and well-suited for mobile applications while delivering exceptional performance. The proposed model is assessed using two benchmark datasets: Plant Disease Expert , consisting of 199,644 images across 58 classes, and PlantVillage , with 54,305 images across 38 classes. Through a rigorous training strategy, Mob-Res demonstrates robust performance, achieving 97.73% average accuracy on the Plant Disease Expert dataset and 99.47% on the PlantVillage dataset. The cross-domain validation rate ( CDVR ) is computed to assess its cross-domain adaptability, with the model showing competitive results compared to other pre-trained models. Additionally, Mob-Res outperforms prominent pre-trained CNN architectures, surpassing ViT-L32 while maintaining a significantly lower parameter count and achieving faster inference times. The proposed model enhances interpretability by utilizing Gradient-weighted Class Activation Mapping ( Grad-CAM ), Grad-CAM++ , and Local Interpretable Model-agnostic Explanations ( LIME ). These techniques provide visual insights into the neural regions influencing the predictions. The experimental results conducted in the current work highlight Mob-Res as a promising solution for automated plant disease detection, supporting large-scale agricultural operations and advancing global food security.

Why it matches plant phenotyping methods植物画像から病害状態を推定するCNNを開発し、複数データセットで性能比較・検証しているため、画像ベースの植物フェノタイピング手法が中心である。

titleA lightweight and explainable CNN model for empowering plant disease diagnosis

A deep learning-based framework for large-scale plant disease detection using big data analytics in precision agriculture →

Abstract Wheat and corn are essential crops for global food security, but wheat yellow rust and the corn northern leaf spot are significant threats. Proper assessment of the severity of the disease is the key to effective control and minimizing crop loss. Traditional methods don’t work effectively, and current deep learning models have problems like focusing too little on severity assessment, only being able to be used for a single crop or disease, and relying on small datasets, all of which make them less reliable in the real world. This paper addresses these issues. It introduces WY-CN-NASNetLarge, a deep-learning model based on the NASNetLarge architecture. The model is trained using transfer learning, fine-tuning, and several datasets, such as Yellow-Rust-19, Corn Disease and Severity (CD&S), and PlantVillage. These help the model work well in a variety of disease conditions. Data augmentation, the AdamW optimizer, dropout training, and mixed precision training enhance performance and prevent overfitting. The model has 97.33% accuracy for classifying disease severity. It is higher than ResNet152v2, InceptionResNetV2, and DenseNet201. This approach is effective and quick for identifying multiple diseases and rating their severity. It can also help manage diseases in agriculture and prevent crop loss.

Why it matches plant phenotyping methods植物病害の重症度を画像等から推定する深層学習手法の開発・比較検証が中心であり、植物の病害状態を直接評価するフェノタイピング研究に該当する。

abstractIt introduces WY-CN-NASNetLarge, a deep-learning model based on the NASNetLarge architecture.

Detecting and Mapping Invasive Species Across Riparian Corridors via Object Detection Approaches in UAV Imagery: An Example of Impatiens glandulifera . →

Riparian zones in the United Kingdom have high species diversity but are prone to anthropogenic changes and alien plant invasions, like Impatiens glandulifera . However, identification can be challenging due to poor accessibility or visibility via tree canopies. UAVs provide a means to access previously inaccessible areas and capture imagery of the area. In this study, a method is introduced to identify the flowers of invasive species ( Impatiens glandulifera ) and map their locations using a computer vision framework and oblique image capture methods. The process includes thresholding images, image masking, blurring, ellipsoid shape search, noise reduction, and contour extraction. Locations are determined using camera parameters, EXIF data, and the average flower size, then converted into vector format for GIS software. This method is wrapped into a single executable program named the semi-automatic thresholding tool (SATT). A validation set of 312 UAV images from the River Elwy, North Wales, showed high precision (79%-96%) and mean average precision (mAP) scores of 73%-86%. This demonstrates that the SATT consistently and correctly identifies Impatiens glandulifera flowers from UAV imagery, making it effective for identifying hotspots and targeting management techniques along riparian corridors. The tool has been wrapped into a single-file executable program with a graphical user interface, enabling nonexperts to use the tool without the need of any software installation. Overall, the tool obtains consistent detection levels of abundance/or flower density across the study site. The tool also does not require an extensive amount of training data, and the intuitive design of the software enables nonexperts to utilize the tool and modify parameter values to adapt it to their needs.

Why it matches plant phenotyping methodsUAV画像から花を検出・抽出し、花の存在位置だけでなく個体群の abundance／flower density を推定する手法と実行可能なツールを開発・検証しており、植物器官形質の取得が中心です。

abstracta method is introduced to identify the flowers of invasive species ( Impatiens glandulifera ) and map their locations using a computer vision framework and oblique image capture methods.

Individual Segmentation of Intertwined Apple Trees in a Row via Prompt Engineering. →

Computer vision is of wide interest to perform the phenotyping of horticultural crops such as apple trees at high throughput. In orchards specially constructed for variety testing or breeding programs, computer vision tools should be able to extract phenotypical information form each tree separately. We focus on segmenting individual apple trees as the main task in this context. Segmenting individual apple trees in dense orchard rows is challenging because of the complexity of outdoor illumination and intertwined branches. Traditional methods rely on supervised learning, which requires a large amount of annotated data. In this study, we explore an alternative approach using prompt engineering with the Segment Anything Model and its variants in a zero-shot setting. Specifically, we first detect the trunk and then position a prompt (five points in a diamond shape) located above the detected trunk to feed to the Segment Anything Model. We evaluate our method on the apple REFPOP, a new large-scale European apple tree dataset and on another publicly available dataset. On these datasets, our trunk detector, which utilizes a trained YOLOv11 model, achieves a good detection rate of 97% based on the prompt located above the detected trunk, achieving a Dice score of 70% without training on the REFPOP dataset and 84% without training on the publicly available dataset.We demonstrate that our method equals or even outperforms purely supervised segmentation approaches or non-prompted foundation models. These results underscore the potential of foundational models guided by well-designed prompts as scalable and annotation-efficient solutions for plant segmentation in complex agricultural environments.

Why it matches plant phenotyping methodsリンゴ樹を個体別に画像分割し、育種・品種試験向けの表現型情報抽出を可能にする手法を開発・評価しており、植物フェノタイピング手法が中心である。

abstractComputer vision is of wide interest to perform the phenotyping of horticultural crops such as apple trees at high throughput.

Plant Disease Detection Using Deep Learning →

The early and accurate detection of plant diseases plays a pivotal role in enhancing crop health and ensuring food security in modern agriculture. Traditional disease diagnosis techniques often rely on visual inspection by experts, which may be subjective, time-consuming, and inaccessible to farmers in remote regions. To overcome these limitations, this project introduces an intelligent plant disease detection system that leverages deep learning and image analysis to identify symptoms directly from leaf images. The proposed methodology begins with the collection of a curated dataset comprising various plant species exhibiting both healthy and diseased conditions. The images undergo preprocessing to enhance quality and ensure consistency, followed by feature extraction using Convolutional Neural Networks (CNNs). Transfer learning is applied to improve model generalization and reduce the training time by utilizing pre-trained models. The system is integrated into a Flask-based web application, enabling users to upload leaf images and receive instant disease diagnoses along with treatment suggestions and suitable fertilizers. Evaluation of the model has shown high classification accuracy across multiple disease classes, affirming its potential to support precision agriculture. The solution is designed to be lightweight, user-friendly, and deployable in real-world agricultural settings, aiming to assist farmers with timely and informed interventions to mitigate crop loss and promote sustainable farming practices.

Why it matches plant phenotyping methods葉画像から植物の病徴を深層学習で抽出・分類するシステム開発が中心であり、植物の病害状態を直接推定するフェノタイピング手法に該当する。

abstractthis project introduces an intelligent plant disease detection system that leverages deep learning and image analysis to identify symptoms directly from leaf images.

Deep Learning Architecture for Tomato Plant Leaf Detection in Images Captured in Complex Outdoor Environments →

The detection of plant constituents is a crucial issue in precision agriculture, as monitoring these enables the automatic analysis of factors such as growth rate, health status, and crop yield. Tomatoes (Solanum sp.) are an economically and nutritionally important crop in Mexico and worldwide, which is why automatic monitoring of these plants is of great interest. Detecting leaves on images of outdoor tomato plants is challenging due to the significant variability in the visual appearance of leaves. Factors like overlapping leaves, variations in lighting, and environmental conditions further complicate the task of detection. This paper proposes modifications to the Yolov11n architecture to improve the detection of tomato leaves in images of complex outdoor environments by incorporating attention modules, transformers, and WIoUv3 loss for bounding box regression. The results show that our proposal led to a 26.75% decrease in the number of parameters and a 7.94% decrease in the number of FLOPs compared with the original version of Yolov11n. Our proposed model outperformed Yolov11n and Yolov12n architectures in recall, F1-measure, and mAP@50 metrics.

Why it matches plant phenotyping methodsトマト葉の画像検出を改善する深層学習アーキテクチャ自体が中心的な技術貢献であり、植物器官の画像ベース取得・推定手法に該当する。

abstractThis paper proposes modifications to the Yolov11n architecture to improve the detection of tomato leaves in images of complex outdoor environments

A fine tuned EfficientNet-B0 convolutional neural network for accurate and efficient classification of apple leaf diseases →

Precise classification and detection of apple diseases are essential for efficient crop management and maximizing yield. This paper presents a fine-tuned EfficientNet-B0 convolutional neural network (CNN) for the automated classification of apple leaf diseases. The model builds upon a pre-trained EfficientNet-B0 base, enhanced through architectural modifications such as the integration of a global max pooling (GMP) layer, dropout, regularization, and full-model fine-tuning. To address class imbalance and improve generalization, the study adopts a holistic training strategy that integrates data augmentation, stratified data splitting, and class weighting, alongside transfer learning. The model is evaluated on the PlantVillage (PV) dataset and a curated Apple PV (APV) dataset and compared against EfficientNet-B0, EfficientNet-B3, Inception-v3, ResNet50, and VGG16 models. The fine-tuned model demonstrates outstanding test accuracies of 99.69% and 99.78% for classifying plant diseases using the APV and PV datasets, respectively. The fine-tuned model outperforms EfficientNet-B0, EfficientNet-B3, and VGG16 on both datasets and shows superior performance compared to Inception-v3 and ResNet-50 on the PV dataset. Both EfficientNet-B0 and the fine-tuned model demonstrate the lowest memory consumption and floating-point operations per second (FLOPs). Also, as compared to the EfficientNet-B0 model, the fine-tuned model achieves an 11% increase in accuracy on the APV dataset and a 49.5% accuracy improvement on the PV dataset, with approximately a 7-8% increase in both memory usage and FLOPs. The fine-tuned model thus emerges as an effective solution for plant leaf disease classification, delivering outstanding accuracy with optimized memory consumption and FLOPs, making it suitable for resource-constrained environments. This study demonstrates that fine-tuned CNN approaches, when combined with transfer learning, advanced data pre-processing, and architectural optimizations, can significantly enhance the accuracy of diseased leaf classification in crops with efficient implementation in limited-resource settings.

Why it matches plant phenotyping methodsリンゴ葉の病害状態を画像から分類するCNN手法の開発・比較評価が研究の中心であり、植物病害表現型の取得・推定に該当する。

abstractThis paper presents a fine-tuned EfficientNet-B0 convolutional neural network (CNN) for the automated classification of apple leaf diseases.

Tomato leaf disease detection method based on improved YOLOv8n. →

With the increasing demand for precision agriculture, automatic detection of tomato leaf diseases has become a critical technological challenge in smart agriculture. Among various diseases, Tomato Yellow Virus Leaf, due to its unique pathological characteristics, presents a particularly challenging identification target. Traditional image recognition methods often fail to meet the high-precision detection requirements for this disease, leading to delayed responses in disease control by farmers, which severely impacts tomato yield and quality. To address this issue, this paper proposes an optimized YOLOv8n algorithm, incorporating a C2f-DynamicConv optimization module. By dynamically adjusting the weights of convolutional kernels, the model can adapt to the characteristics of different input data, thereby enhancing its ability to represent diverse features. Additionally, we introduce the SimAM attention mechanism, which enhances the model's focus on key areas by weighting the feature map, significantly improving the accuracy of disease detection while filtering out irrelevant features and enhancing sensitivity. During the upsampling process, we adopt the Dysample upsampling operator, optimizing the quality of feature map reconstruction and improving detection resolution through a refined upsampling strategy. To better address the bounding box regression problem in object detection, we incorporate the GIoU loss function. Compared to traditional loss functions, GIoU performs excellently in handling bounding box overlap and positional accuracy, further improving the model's detection performance. Experimental results show that the improved model achieves an average precision of 81.8%, precision of 77.1%, and recall of 77.4%. Compared to existing methods, our approach shows significant advantages in detection accuracy, localization precision, and model computational efficiency, achieving improved detection performance on the tomato leaf disease dataset.

Why it matches plant phenotyping methodsトマト葉の病徴を画像から検出する改良YOLOv8n手法を開発し、精度・再現率・位置特定性能を評価しており、植物の病害状態の取得が研究の中心である。

abstractExperimental results show that the improved model achieves an average precision of 81.8%, precision of 77.1%, and recall of 77.4%.

Detection of Rice and Corn Plant Leaf Disease Using Invariants of Deep Learning Models and Edge Perspective →

Rice and corn hold significant importance due to their daily consumption worldwide. Naked-eye observations are not accurate. Therefore, we need an autonomous system that can accurately detect and classify diseases in both plants. We trained and validated publicly available datasets in three deep convolutional neural network (DCNN)-based deep learning models using different learning rates and found that the lowest learning rate was the most effective in achieving the highest accuracy. We added a new dense layer to the known DCNN-based deep learning models and achieved improved accuracy. The best results were observed when our invariants of the InceptionV3, ResNet152, and MobileNetV2 deep learning models were used on corn plant leaves (98.09%, 98.51%, and 89.73%, respectively). These models also performed well on rice plant leaves (98.51%, 93.59%, and 98.57%, respectively). Because InceptionV3 performed well for both plants, we implemented it in NVIDIA Jetson Nano as an end device for the detection and classification of diseases from both plant leaves. Received: 4 January 2025 | Revised: 26 May 2025 | Accepted: 13 June 2025 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement The data that support the findings of this study are openly available in Kaggle at https://www.kaggle.com/datasets/vbookshelf/rice-leaf-diseases and https://www.kaggle.com/datasets/smaranjitghose/corn-or-maize-leaf-disease-dataset. Author Contribution Statement Zubair Saeed: Conceptualization, Methodology, Software, Validation, Resources, Data curation, Writing – original draft, Project administration. Uzma Nawaz: Validation, Formal analysis, Writing – review & editing. Ali Raza: Conceptualization, Validation, Formal analysis, Writing – review & editing, Visualization. Kamran Javed: Validation, Formal analysis, Investigation, Writing – review & editing, Visualization, Supervision.

Why it matches plant phenotyping methods植物葉の病徴を画像から検出・分類する深層学習手法を開発・検証し、エッジデバイスにも実装しており、植物状態の取得・推定が研究の中心です。

abstractWe trained and validated publicly available datasets in three deep convolutional neural network (DCNN)-based deep learning models using different learning rates and found that the lowest learning rate was the most effective in achieving the highest accuracy.

Tomato Multi-Angle Multi-Pose Dataset for Fine-Grained Phenotyping →

Observer bias and inconsistencies in traditional plant phenotyping methods limit the accuracy and reproducibility of fine-grained plant analysis. To overcome these challenges, we developed TomatoMAP, a comprehensive dataset for Solanum lycopersicum using an Internet of Things (IoT) based imaging system with standardized data acquisition protocols. Our dataset contains 64,464 RGB images that capture 12 different plant poses from four camera elevation angles. Each image includes manually annotated bounding boxes for seven regions of interest (ROIs), including leaves, panicle, batch of flowers, batch of fruits, axillary shoot, shoot and whole plant area, along with 50 fine-grained growth stage classifications based on the BBCH scale. Additionally, we provide 3,616 high-resolution image subset with pixel-wise semantic and instance segmentation annotations for fine-grained phenotyping. We validated our dataset using a cascading model deep learning framework combining MobileNetv3 for classification, YOLOv11 for object detection, and MaskRCNN for segmentation. Through AI vs. Human analysis involving five domain experts, we demonstrate that the models trained on our dataset achieve accuracy and speed comparable to the experts. Cohen's Kappa and inter-rater agreement heatmap confirm the reliability of automated fine-grained phenotyping using our approach.

Why it matches plant phenotyping methods植物の多視点画像取得、アノテーション付きデータセット、深層学習による分類・検出・セグメンテーションを中心に開発・検証した、明確な植物フェノタイピング手法研究です。

abstractwe developed TomatoMAP, a comprehensive dataset for Solanum lycopersicum using an Internet of Things (IoT) based imaging system with standardized data acquisition protocols.

A Lightweight and Efficient Plant Disease Detection Method Integrating Knowledge Distillation and Dual-Scale Weighted Convolutions →

Plant diseases significantly undermine agricultural productivity. This study introduces an improved YOLOv10n model named WD-YOLO (Weighted and Double-scale YOLO), an advanced architecture for efficient plant disease detection. The PlantDoc dataset was initially enhanced using data augmentation techniques. Subsequently, we developed the DSConv module—a novel convolutional structure employing double-scale weighted convolutions that dynamically adjust to different scale perceptions and optimize attention allocation. This module replaces the conventional Conv module in YOLOv10. Furthermore, the WTConcat module was introduced, dynamically merging weighted concatenation with a channel attention mechanism to replace the Concat module in YOLOv10. The training of WD-YOLO incorporated knowledge distillation techniques using YOLOv10l as a teacher model to refine and compress the architectural learning. Empirical results reveal that WD-YOLO achieved an mAP50 of 65.4%, outperforming YOLOv10n by 9.1% without data augmentation and YOLOv10l by 2.3%, despite having significantly fewer parameters (9.3 times less than YOLOv10l), demonstrating substantial gains in detection efficiency and model compactness.

Why it matches plant phenotyping methods植物病害を画像から検出するモデルの新規構造（DSConv、WTConcat、知識蒸留）を開発・評価しており、植物の病害状態を推定する方法が研究の中心である。

abstractThis study introduces an improved YOLOv10n model named WD-YOLO (Weighted and Double-scale YOLO), an advanced architecture for efficient plant disease detection.

Leaf Doctor: An Advanced Plant Disease Detection Web Application →

Leaf Doctor is an advanced plant disease detection web application that assists farmers and agricultural researchers in diagnosing plant diseases through image analysis. Using machine learning algorithms, the system analyzes leaf images to identify signs of disease, providing users with accurate and timely insights. The platform enhances agricultural efficiency by offering real-time disease detection, reducing crop loss, and promoting sustainable farming practices. As a cloud-based solution, Leaf Doctor is accessible from multiple devices, ensuring widespread usability for farmers and agronomists. Key Words: Plant Disease Detection, Gated Recurrent Unit, Leaf Image Analysis, Streamlit, Real Time Diagnosis, Sustainable Farming.

Why it matches plant phenotyping methods葉画像から植物病徴を機械学習で検出するWebアプリケーションが研究の中心であり、植物の病害状態を観測・推定するフェノタイピング手法に該当する。

abstractLeaf Doctor is an advanced plant disease detection web application that assists farmers and agricultural researchers in diagnosing plant diseases through image analysis.

A study on an efficient citrus Huanglong disease detection algorithm based on three-channel aggregated attention. →

Background Aiming at the problems of complex and diverse field symptoms of citrus Huanglong disease (HLB), low efficiency and insufficient recognition accuracy of traditional detection methods, this study proposes an efficient detection algorithm based on improved You Only Look Once (YOLO)v8. Methods Firstly, a new character to float (C2f) Attention inverse residual moving block (IRMB) module is designed, which significantly enhances the model's sensitivity to tiny disease features while reducing the number of parameters by fusing the lightweight IRMB with the adaptive attention gating mechanism, and solves the problem of losing key texture information due to downsampling in the traditional C2f module. Secondly, the three-channel aggregated attention module Powerneck is proposed in the Neck section, which realizes efficient cross-scale feature interactions, effectively suppresses background noise interference, and improves robustness in complex field scenes through SimFusion_4in feature alignment, information fusion module (IFM) global context fusion, and Power channel dynamic weighting strategy. In addition, the detection head design is optimized by structural reparameterization technique to further accelerate the inference process. Results The experimental results show that on the citrus dataset containing 12 diseases and two health states, the mAP50 of this model reaches 97% and the accuracy is 91.5%, which is 1.1% and 1.2% higher than that of the original YOLOv8, respectively, and the inference speed is improved by 14.6% to 370 frames per second (FPS). Comparison of the different models shows that the C2f Attention IRMB, through the mechanism of dual attention The comparison of different models shows that C2f Attention IRMB strengthens the feature expression ability through the dual-attention mechanism, and the Powerneck module reduces redundant computation through dynamic channel pruning, and the two synergistically optimize the model performance significantly. Compared with mainstream models such as YOLOv5m and YOLOv7x, this method is more advantageous in the balance of accuracy and speed, and can meet the demand of real-time detection in the field. Discussion The algorithm provides an efficient tool for early and accurate identification of citrus Huanglong disease, which is of great practical significance for reducing pesticide misuse and improving the efficiency of orchard management, and also provides new ideas for the design of lightweight target detection models in agricultural scenarios.

Why it matches plant phenotyping methods柑橘HLBの植物症状を画像から検出するYOLOv8改良アルゴリズムを開発・評価しており、病害状態の推定手法が研究の中心である。

abstractthis study proposes an efficient detection algorithm based on improved You Only Look Once (YOLO)v8.

Detection of Citrus Huanglongbing in Natural Field Conditions Using an Enhanced YOLO11 Framework →

Citrus Huanglongbing (HLB) is one of the most devastating diseases in the global citrus industry, but its early detection under complex field conditions remains a major challenge. Existing methods often suffer from insufficient dataset diversity and poor generalization, and struggle to accurately detect subtle early-stage lesions and multiple HLB symptoms in natural backgrounds. To address these issues, we propose an enhanced YOLO11-based framework, DCH-YOLO11. We constructed a multi-symptom HLB leaf dataset (MS-HLBD) containing 9219 annotated images across five classes: Healthy (1862), HLB blotchy mottling (2040), HLB Zinc deficiency (1988), HLB yellowing (1768), and Canker (1561), collected under diverse field conditions. To improve detection performance, the DCH-YOLO11 framework incorporates three novel modules: the C3k2 Dynamic Feature Fusion (C3k2_DFF) module, which enhances early and subtle lesion detection through dynamic feature fusion; the C2PSA Context Anchor Attention (C2PSA_CAA) module, which leverages context anchor attention to strengthen feature extraction in complex vein regions; and the High-efficiency Dynamic Feature Pyramid Network (HDFPN) module, which optimizes multi-scale feature interaction to boost detection accuracy across different object sizes. On the MS-HLBD dataset, DCH-YOLO11 achieved a precision of 91.6%, recall of 87.1%, F1-score of 89.3, and mAP50 of 93.1%, surpassing Faster R-CNN, SSD, RT-DETR, YOLOv7-tiny, YOLOv8n, YOLOv9-tiny, YOLOv10n, YOLO11n, and YOLOv12n by 13.6%, 8.8%, 5.3%, 3.2%, 2.0%, 1.6%, 2.6%, 1.8%, and 1.6% in mAP50, respectively. On a publicly available citrus HLB dataset, DCH-YOLO11 achieved a precision of 82.7%, recall of 81.8%, F1-score of 82.2, and mAP50 of 89.4%, with mAP50 improvements of 8.9%, 4.0%, 3.8%, 3.2%, 4.7%, 3.2%, and 3.4% over RT-DETR, YOLOv7-tiny, YOLOv8n, YOLOv9-tiny, YOLOv10n, YOLO11n, and YOLOv12n, respectively. These results demonstrate that DCH-YOLO11 achieves both state-of-the-art accuracy and excellent generalization, highlighting its strong potential for robust and practical citrus HLB detection in real-world applications.

Why it matches plant phenotyping methods柑橘葉のHLB症状を画像から検出するYOLOベースの表現型取得手法を開発し、専用データセットと公開データセットで性能検証しているため、植物病害表現型の方法研究として中心的である。

abstractwe propose an enhanced YOLO11-based framework, DCH-YOLO11.

Mamba-YOLO-ML: A State-Space Model-Based Approach for Mulberry Leaf Disease Detection. →

Mulberry ( Morus spp.), as an economically significant crop in sericulture and medicinal applications, faces severe threats to leaf yield and quality from pest and disease infestations. Traditional detection methods relying on chemical pesticides and manual observation prove inefficient and unsustainable. Although computer vision and deep learning technologies offer new solutions, existing models exhibit limitations in natural environments, including low recognition rates for small targets, insufficient computational efficiency, poor adaptability to occlusions, and inability to accurately identify structural features such as leaf veins. We propose Mamba-YOLO-ML, an optimized model addressing three key challenges in vision-based detection: Phase-Modular Design (PMSS) with dual blocks enhancing multi-scale feature representation and SSM selective mechanisms and Mamba Block, Haar wavelet downsampling preserving critical texture details, and Normalized Wasserstein Distance loss improving small-target robustness. Visualization analysis of the detection performance on the test set using GradCAM revealed that the enhanced Mamba-YOLO-ML model demonstrates earlier and more effective focus on characteristic regions of different diseases compared with its predecessor. The improved model achieved superior detection accuracy with 78.2% mAP50 and 59.9% mAP50:95, outperforming YOLO variants and comparable Transformer-based models, establishing new state-of-the-art performance. Its lightweight architecture (5.6 million parameters, 13.4 GFLOPS) maintains compatibility with embedded devices, enabling real-time field deployment. This study provides an extensible technical solution for precision agriculture, facilitating sustainable mulberry cultivation through efficient pest and disease management.

Why it matches plant phenotyping methods桑葉の病害状態を画像から検出する深層学習モデルを開発・評価しており、植物病害表現型の取得手法が研究の中心である。

abstractWe propose Mamba-YOLO-ML, an optimized model addressing three key challenges in vision-based detection

StomaYOLO: A Lightweight Maize Phenotypic Stomatal Cell Detector Based on Multi-Task Training. →

L.), a vital global food crop, relies on its stomatal structure for regulating photosynthesis and responding to drought. Conventional manual stomatal detection methods are inefficient, subjective, and inadequate for high-throughput plant phenotyping research. To address this, we curated a dataset of over 1500 maize leaf epidermal stomata images and developed a novel lightweight detection model, StomaYOLO, tailored for small stomatal targets and subtle features in microscopic images. Leveraging the YOLOv11 framework, StomaYOLO integrates the Small Object Detection layer P2, the dynamic convolution module, and exploits large-scale epidermal cell features to enhance stomatal recognition through auxiliary training. Our model achieved a remarkable 91.8% mean average precision (mAP) and 98.5% precision, surpassing numerous mainstream detection models while maintaining computational efficiency. Ablation and comparative analyses demonstrated that the Small Object Detection layer, dynamic convolutional module, multi-task training, and knowledge distillation strategies substantially enhanced detection performance. Integrating all four strategies yielded a nearly 9% mAP improvement over the baseline model, with computational complexity under 8.4 GFLOPS. Our findings underscore the superior detection capabilities of StomaYOLO compared to existing methods, offering a cost-effective solution that is suitable for practical implementation. This study presents a valuable tool for maize stomatal phenotyping, supporting crop breeding and smart agriculture advancements.

Why it matches plant phenotyping methodsトウモロコシの気孔を画像から検出するモデルとデータセットを開発・評価しており、植物フェノタイピング手法が研究の中心です。

abstractwe curated a dataset of over 1500 maize leaf epidermal stomata images and developed a novel lightweight detection model, StomaYOLO

Title: KymoTip: High-throughput Characterization of Tip-growth Dynamics in Plant Cells →

Summary Live imaging data analysis often requires an objective, local, and accurate way of quantification of cell dynamics. In the research field of polarized tip-growth, the cell fluctuations and/or fluctuations in tip position and growth direction hampers automated analyses of huge amounts of imaging sequences. The fluctuated nature in data makes it unclear how cell shape and growth are linked to intracellular events that could be the actual driving force of cell growth. To overcome these difficulties, we developed a powerful and user-friendly tool called KymoTip with an available format. In this software, novel functions such as coordinate normalization, tip-bottom detection, and signal kymograph were implemented. We confirmed that not only plasma membrane-labeled fluorescent images, but also images such as bright-field and cortical microtubule markers —so long as the cell contours can be identified— are amenable to KymoTip. Furthermore, by combining markers for cell contours with those that visualize intracellular structures, it becomes possible to quantitatively analyze various intracellular events, such as nuclear migration and calcium wave, in conjunction with cellular growth dynamics. Since KymoTip can be handled by non-specialist, it is expected to promote understanding of what happens at the sub- and cellular level with high throughput outcomes. Significance statement Faced with fluctuations in cell coordinates and cell tip positions, position correction of live imaging data and accurate detection of tip position are key challenges in plant developmental biology. We solved them with a powerful and user-friendly tool, KymoTip, that can realize cell position correction, cell tip detection with cell centerline, and quantification of intracellular events.

Why it matches plant phenotyping methods植物細胞のライブイメージから細胞形状・先端位置・成長動態を定量化する解析ソフトウェアを開発しており、植物フェノタイピング手法が中心です。

abstractwe developed a powerful and user-friendly tool called KymoTip

On the Minimum Dataset Requirements for Fine-Tuning an Object Detector for Arable Crop Plant Counting: A Case Study on Maize Seedlings →

Object detection is essential for precision agriculture applications like automated plant counting, but the minimum dataset requirements for effective model deployment remain poorly understood for arable crop seedling detection on orthomosaics. This study investigated how much annotated data is required to achieve standard counting accuracy (R2 = 0.85) for maize seedlings across different object detection approaches. We systematically evaluated traditional deep learning models requiring many training examples (YOLOv5, YOLOv8, YOLO11, RT-DETR), newer approaches requiring few examples (CD-ViTO), and methods requiring zero labeled examples (OWLv2) using drone-captured orthomosaic RGB imagery. We also implemented a handcrafted computer graphics algorithm as baseline. Models were tested with varying training sources (in-domain vs. out-of-distribution data), training dataset sizes (10–150 images), and annotation quality levels (10–100%). Our results demonstrate that no model trained on out-of-distribution data achieved acceptable performance, regardless of dataset size. In contrast, models trained on in-domain data reached the benchmark with as few as 60–130 annotated images, depending on architecture. Transformer-based models (RT-DETR) required significantly fewer samples (60) than CNN-based models (110–130), though they showed different tolerances to annotation quality reduction. Models maintained acceptable performance with only 65–90% of original annotation quality. Despite recent advances, neither few-shot nor zero-shot approaches met minimum performance requirements for precision agriculture deployment. These findings provide practical guidance for developing maize seedling detection systems, demonstrating that successful deployment requires in-domain training data, with minimum dataset requirements varying by model architecture.

Why it matches plant phenotyping methodsトウモロコシ幼苗の個体数という植物形質を画像から推定する物体検出手法について、複数モデル、データ量、アノテーション品質を系統的に比較・評価しており、手法の性能検証が中心である。

abstractThis study investigated how much annotated data is required to achieve standard counting accuracy (R2 = 0.85) for maize seedlings across different object detection approaches.

A Hybrid Deep Learning Approach for Cotton Plant Disease Detection Using BERT-ResNet-PSO →

Cotton is one of the most valuable non-food agricultural products in the world. However, cotton production is often hampered by the invasion of disease. In most cases, these plant diseases are a result of insect or pest infestations, which can have a significant impact on production if not addressed promptly. It is, therefore, crucial to accurately identify leaf diseases in cotton plants to prevent any negative effects on yield. This paper presents a hybrid deep learning approach based on Bidirectional Encoder Representations from Transformers with Residual network and particle swarm optimization (BERT-ResNet-PSO) for detecting cotton plant diseases. This approach starts with image pre-processing, which they pass to a BERT-like encoder after linearly embedding the image patches. It results in segregating disease regions. Then, the output of the encoded feature is passed to ResNet-based architecture for feature extraction and further optimized by PSO to increase the classification accuracy. The approach is tested on a cotton dataset from the Plant Village dataset, where the experimental results show the effectiveness of this hybrid deep learning approach, achieving an accuracy of 98.5%, precision of 98.2% and recall of 98.7% compared to the existing deep learning approaches such as ResNet50, VGG19, InceptionV3, and ResNet152V2. This study shows that the hybrid deep learning approach is capable of dealing with the cotton plant disease detection problem effectively. This study suggests that the proposed approach is beneficial to help avoid crop losses on a large scale and support effective farming management practices.

Why it matches plant phenotyping methods葉画像から綿花の病害領域を抽出・分類する深層学習手法を開発し、既存手法と比較評価しており、植物病害状態の表現型取得が中心である。

abstractThis paper presents a hybrid deep learning approach based on Bidirectional Encoder Representations from Transformers with Residual network and particle swarm optimization (BERT-ResNet-PSO) for detecting cotton plant diseases.

MatchPlant: An Open-Source Pipeline for UAV-Based Single-Plant Detection and Data Extraction →

Accurate identification of individual plants from unmanned aerial vehicle (UAV) images is essential for advancing high-throughput phenotyping and supporting data-driven decision-making in plant breeding. This study presents MatchPlant, a modular, graphical user interface-supported, open-source Python pipeline for UAV-based single-plant detection and geospatial trait extraction. MatchPlant enables end-to-end workflows by integrating UAV image processing, user-guided annotation, Convolutional Neural Network model training for object detection, forward projection of bounding boxes onto an orthomosaic, and shapefile generation for spatial phenotypic analysis. In an early-season maize case study, MatchPlant achieved reliable detection performance (validation AP: 89.6%, test AP: 85.9%) and effectively projected bounding boxes, covering 89.8% of manually annotated boxes with 87.5% of projections achieving an Intersection over Union (IoU) greater than 0.5. Trait values extracted from predicted bounding instances showed high agreement with manual annotations (r = 0.87-0.97, IoU >= 0.4). Detection outputs were reused across time points to extract plant height and Normalized Difference Vegetation Index with minimal additional annotation, facilitating efficient temporal phenotyping. By combining modular design, reproducibility, and geospatial precision, MatchPlant offers a scalable framework for UAV-based plant-level analysis with broad applicability in agricultural and environmental monitoring.

Why it matches plant phenotyping methodsUAV画像から個体検出・地理空間的形質抽出を行うオープンソース基盤の開発と性能検証が中心であり、植物形質（草丈・NDVI）を抽出する再利用可能なワークフローを提供している。

abstractThis study presents MatchPlant, a modular, graphical user interface-supported, open-source Python pipeline for UAV-based single-plant detection and geospatial trait extraction.

Machine learning assisted analysis of rice flower opening times using a low-cost time-lapse camera. →

Flower opening time (FOT) is a key trait for successful reproduction and reproductive isolation. In crop science, FOT is critical for stress avoidance and efficient breeding practices. This study developed a system for the automatic detection of rice flower openings and FOT estimation by integrating a low-cost time-lapse camera with machine learning technology. This approach enabled high-resolution monitoring of flowering dynamics in two cultivars: the japonica cultivar Taichung 65 (T65) and the indica cultivar IR24. The system accurately identified regions containing open flowers, and the estimated FOTs varied within a 3-h range, with a root mean square error of approximately 30 min compared to manual detection. A significant difference in estimated FOTs between IR24 and T65 demonstrated the system's potential for genetic screening applications. FOT of both cultivars exhibited a significant negative correlation with daily mean temperature. Notably, a temperature-sensitive period was identified in the morning, suggesting that temperature influences not only flower opening but also preceding physiological processes such as panicle and spikelet development. This study presents a novel approach to investigating FOT dynamics in rice and provides insights into the interaction between environmental factors and internal regulatory mechanisms governing this critical reproductive trait.

Why it matches plant phenotyping methods低コストタイムラプスカメラと機械学習によるイネの開花時刻という植物形質の自動検出・推定システムを開発し、手動検出との誤差で検証しているため、方法が中心的です。

abstractThis study developed a system for the automatic detection of rice flower openings and FOT estimation by integrating a low-cost time-lapse camera with machine learning technology.

Iterative segmentation and classification for enhanced crop disease diagnosis using optimized hybrid U-Nets model →

The major challenges that the agricultural sector faces are that with the kind of methodologies that exist, gross limitations may occur to the exact diagnosis of crop diseases. They are unable to achieve correct precision in disease classification, relatively lower accuracy, and delayed response time—all these obstacles result in a deficiency in effectual disease management and control. Our research proposes a new framework instigated and developed to improve crop disease detection and classification by multifaceted analysis. In the core of our methodology is the implementation of adaptive anisotropic diffusion for the denoising of obtained agro images, therefore making it a step towards assurance in data quality. Along with this is the use of a Fuzzy U-Net++ model for image segmentation, whereby fuzzy decisions in generously instill an increase in performance for image segmentation. Feature selection itself is innovated by the introduction of the Moving Gorilla Remora Algorithm (MGRA) combined with convolutional operations, setting a new benchmark in the selection of optimal features pertaining to disease identification operations. To further refine this model, classification is adeptly handled by a process inspired by the LeNet architecture, significantly improving identification against various diseases. Our approach’s performance is therefore strongly assessed through a number of renowned datasets, such as PlantVillage and PlantDoc, on which test metrics show superior performance: 8.5% improvement in disease classification precision, 8.3% higher accuracy, 9.4% improved recall, with a reduction in time delay by 4.5%, area under the curve (AUC) increasing by 5.9%, a 6.5% improvement in specificity, far ahead of other methods. This work not only sets new standards in crop disease analysis but also opens possibilities for the preemptive measures to come in agricultural health, promising a future where crop management is more effective and efficient. Our results thus have implications that reach beyond the immediate benefits accruable from improved diagnosis of diseases. It is a harbinger of a new era in agricultural technology where precision, accuracy, and timeliness will meet to enhance crop resilience and yield.

Why it matches plant phenotyping methods植物画像から病害症状をセグメンテーション・分類する計算手法が研究の中心であり、PlantVillageおよびPlantDocで性能評価も行っているため、植物フェノタイピング手法として採用。

abstractOur research proposes a new framework instigated and developed to improve crop disease detection and classification by multifaceted analysis.

FRPNet: A Lightweight Multi-Altitude Field Rice Panicle Detection and Counting Network Based on Unmanned Aerial Vehicle Images →

Rice panicle detection is a key technology for improving rice yield and agricultural management levels. Traditional manual counting methods are labor-intensive and inefficient, making them unsuitable for large-scale farmlands. This paper proposes FRPNet, a novel lightweight convolutional neural network optimized for multi-altitude rice panicle detection in UAV images. The architecture integrates three core innovations: a CSP-ScConv backbone with self-calibrating convolutions for efficient multi-scale feature extraction; a Feature Pyramid Shared Convolution (FPSC) module that replaces pooling with multi-branch dilated convolutions to preserve fine-grained spatial information; and a Dynamic Bidirectional Feature Pyramid Network (DynamicBiFPN) employing input-adaptive kernels to optimize cross-scale feature fusion. The model was trained and evaluated on the open-access Dense Rice Panicle Detection (DRPD) dataset, which comprises UAV images captured at 7 m, 12 m, and 20 m altitudes. Experimental results demonstrate that our method significantly outperforms existing advanced models, achieving an AP50 of 0.8931 and an F2 score of 0.8377 on the test set. While ensuring model accuracy, the parameters of the proposed model decreased by 42.87% and the GFLOPs by 48.95% compared to Panicle-AI. Grad-CAM visualizations reveal that FRPNet exhibits superior background noise suppression in 20 m altitude images compared to mainstream models. This work establishes an accuracy-efficiency balanced solution for UAV-based field phenotyping.

Why it matches plant phenotyping methodsUAV画像からイネ穂の検出・計数という植物形質を抽出する軽量深層学習手法を開発し、公開データセットで評価しており、表現型取得手法が中心である。

abstractThis paper proposes FRPNet, a novel lightweight convolutional neural network optimized for multi-altitude rice panicle detection in UAV images.

OpenPheno: an open-access, user-friendly, and smartphone-based software platform for instant plant phenotyping. →

BACKGROUND: Plant phenotyping has become increasingly important for advancing plant science, agriculture, and biotechnology. Classic manual methods are labor-intensive and time-consuming, while existing computational tools often require advanced coding skills, high-performance hardware, or PC-based environments, making them inaccessible to non-experts, to resource-constrained users, and to field technicians. RESULTS: To respond to these challenges, we introduce OpenPheno, an open-access, user-friendly, and smartphone-based platform encapsulated within a WeChat Mini-Program for instant plant phenotyping. The platform is designed for ease of use, enabling users to phenotype plant traits quickly and efficiently with only a smartphone at hand. We currently instantiate the use of the platform with tools such as SeedPheno, WheatHeadPheno, LeafAnglePheno, SpikeletPheno, CanopyPheno, TomatoPheno, and CornPheno; each offering specific functionalities such as seed size and count analysis, wheat head detection, leaf angle measurement, spikelet counting, canopy structure analysis, and tomato fruit measurement. In particular, OpenPheno allows developers to contribute new algorithmic tools, further expanding its capabilities to continuously facilitate the plant phenotyping community. CONCLUSIONS: By leveraging cloud computing and a widely accessible interface, OpenPheno democratizes plant phenotyping, making advanced tools available to a broader audience, including plant scientists, breeders, and even amateurs. It can function as a role in AI-driven breeding by providing the necessary data for genotype-phenotype analysis, thereby accelerating breeding programs. Its integration with smartphones also positions OpenPheno as a powerful tool in the growing field of mobile-based agricultural technologies, paving the way for more efficient, scalable, and accessible agricultural research and breeding.

Why it matches plant phenotyping methodsスマートフォンで植物形質を取得・解析するソフトウェアプラットフォームの開発が中心であり、複数の具体的な表現型解析ツールを提供している。

abstractwe introduce OpenPheno, an open-access, user-friendly, and smartphone-based platform encapsulated within a WeChat Mini-Program for instant plant phenotyping.

Smart Diagnosis: Early Detection and Management of Plant Diseases →

Plant diseases threaten global food security and cause significant financial losses in agriculture. Early detection and precise diagnosis are critical for effective disease management. This study explores a novel approach to plant disease identification using the You Only Look Once (YOLOv12) algorithm combined with few-shot learning techniques. By leveraging a limited dataset, the model is trained to classify citrus plant leaf images into four categories: healthy, greening, black spot, and canker. The proposed system enhances disease detection efficiency, enabling farmers to take timely preventive measures. Our approach demonstrates the potential of few-shot learning in agricultural disease diagnosis, reducing the need for extensive labeled datasets while maintaining high accuracy.

Why it matches plant phenotyping methods柑橘葉画像から健全・greening・黒点病・かんきつかいよう病を分類するYOLOv12とfew-shot learning手法が研究の中心であり、植物病害状態の画像ベース推定に該当する。

abstractThis study explores a novel approach to plant disease identification using the You Only Look Once (YOLOv12) algorithm combined with few-shot learning techniques.

Enhancing the dataset of CycleGAN-M and YOLOv8s-KEF for identifying apple leaf diseases. →

Accurate diagnosis of apple diseases is vital for tree health, yield improvement, and minimizing economic losses. This study introduces a deep learning-based model to tackle issues like limited datasets, small sample sizes, and low recognition accuracy in detecting apple leaf diseases. The approach begins with enhancing the CycleGAN-M network using a multi-scale attention mechanism to generate synthetic samples, improving model robustness and generalization by mitigating imbalances in disease-type representation. Next, an improved YOLOv8s-KEF model is introduced to overcome limitations in feature extraction, particularly for small lesions and complex textures in natural environments. The model's backbone replaces the standard C2f structure with C2f-KanConv, significantly enhancing disease recognition capabilities. Additionally, we optimize the detection head with Efficient Multi-Scale Convolution (EMS-Conv), improving the model's ability to detect small targets while maintaining robustness and generalization across diverse disease types and conditions. Incorporating Focal-EIoU further reduces missed and false detections, enhancing overall accuracy. The experiment results demonstrate that the YOLOv8s-KEF model achieves 95.0% in accuracy, 93.1% in recall, 95.8% in precision, and an F1-score of 94.5%. Compared to the original YOLOv8s model, the proposed model improves accuracy by 7.2%, precision by 6.5%, and F1-score by 5.0%, with only a modest 6MB increase in model size. Furthermore, compared to Faster RCNN, ResNet50, SSD, YOLOv3-tiny, YOLOv6, YOLOv9s, and YOLOv10m, our model demonstrates substantial improvements, with up to 30.2% higher precision and 18.0% greater accuracy. This study used CycleGAN-M and YOLOv8s-KEF methods to enhance the detection capability of apple leaf diseases.

Why it matches plant phenotyping methodsリンゴ葉の病斑・病害を画像から検出する深層学習モデルの改良と性能比較が研究の中心であり、植物の病害状態を直接推定するフェノタイピング手法に該当する。

abstractThis study introduces a deep learning-based model to tackle issues like limited datasets, small sample sizes, and low recognition accuracy in detecting apple leaf diseases.

An efficient plant disease detection using transfer learning approach. →

Plant diseases pose significant challenges to farmers and the agricultural sector at large. However, early detection of plant diseases is crucial to mitigating their effects and preventing widespread damage, as outbreaks can severely impact the productivity and quality of crops. With advancements in technology, there are increasing opportunities for automating the monitoring and detection of disease outbreaks in plants. This study proposed a system designed to identify and monitor plant diseases using a transfer learning approach. Specifically, the study utilizes YOLOv7 and YOLOv8, two state-of-the-art models in the field of object detection. By fine-tuning these models on a dataset of plant leaf images, the system is able to accurately detect the presence of Bacteria, Fungi and Viral diseases such as Powdery Mildew, Angular Leaf Spot, Early blight and Tomato mosaic virus. The model's performance was evaluated using several metrics, including mean Average Precision (mAP), F1-score, Precision, and Recall, yielding values of 91.05, 89.40, 91.22, and 87.66, respectively. The result demonstrates the superior effectiveness and efficiency of YOLOv8 compared to other object detection methods, highlighting its potential for use in modern agricultural practices. The approach provides a scalable, automated solution for early any plant disease detection, contributing to enhanced crop yield, reduced reliance on manual monitoring, and supporting sustainable agricultural practices.

Why it matches plant phenotyping methods植物葉画像から病害の有無・種類を推定する画像ベースの病害表現型計測法をYOLOモデルで開発・評価しており、手法が研究の中心です。

abstractThis study proposed a system designed to identify and monitor plant diseases using a transfer learning approach.

BGM-YOLO: An accurate and efficient detector for detecting plant disease. →

Given the complexity of crop growth environments in nature, where leaf backgrounds often include soil, weeds, and other plants, along with variable lighting conditions, and considering the small size of leaf spots and the wide variety of crop diseases with significant scale differences, this paper proposes a new BGM-YOLO model structure aimed at improving accuracy and inference speed. First, the GSBottleneck module is utilized to enhance the C2f module of the YOLOv8n model, leading to the introduction of the GSC2f module, which reduces computational costs and increases inference efficiency. Next, the model incorporates a multiscale bitemporal fusion module (BFM) to increase the effectiveness and robustness of feature fusion across different levels. Finally, we developed a median-enhanced spatial and channel attention block (MECS) that combines both channel and spatial attention mechanisms, effectively improving the capture and fusion of small-scale features. The experimental results demonstrate that the BGM-YOLO model achieves a 3.9% improvement in the mean average precision (mAP) over the original model. In crop disease detection tasks, the BGM-YOLO model has higher detection accuracy and a lower false negative rate, confirming its practical value in complex application scenarios.

Why it matches plant phenotyping methods植物の病斑・病害状態を画像から検出するYOLOベース手法を開発し、精度と推論速度を評価しており、植物病害フェノタイピング手法が中心です。

abstractthis paper proposes a new BGM-YOLO model structure aimed at improving accuracy and inference speed.

FieldDino: Rapid In-Field Stomatal Anatomy and Physiology Phenotyping. →

ABSTRACT Stomatal anatomy and physiology define CO 2 availability for photosynthesis and regulate plant water use. Despite being key drivers of yield and dynamic responsiveness to abiotic stresses, conventional measurement techniques of stomatal traits are laborious and slow, limiting adoption in plant breeding. Advances in instrumentation and data analyses present an opportunity to screen stomatal traits at scales relevant to plant breeding. We present a high‐throughput robust field‐based phenotyping approach, FieldDino, for screening stomatal physiology and anatomy. The method allows measurements to be collected in p @0.5 of 97.1% for stomatal detection. When validated in large field trials of 200 wheat genotypes under two irrigation treatments, FieldDino captured wide diversity in stomatal traits. FieldDino enables stomatal data collection and analysis at unprecedented scales in the field. This will advance research on stomatal biology and accelerate the incorporation of stomatal traits into plant breeding programs for resilience to abiotic stress.

Why it matches plant phenotyping methodsFieldDinoは、圃場で気孔の生理・解剖形質を高スループットに取得するフェノタイピング手法として開発され、大規模圃場試験で検証されているため含める。

abstractWe present a high‐throughput robust field‐based phenotyping approach, FieldDino, for screening stomatal physiology and anatomy.

High-throughput iNaturalist image analysis reveals flower color divergence in Monarda fistulosa →

Characterizing patterns of trait variation across widespread species is a fundamental goal of natural history. Here we create a pipeline to analyze a large community science dataset and test hypothesized flower color divergence across the range of a widespread wildflower. Monarda fistulosa is a North American perennial that produces showy lavender inflorescences. Although previous literature suggests that the flowers of western M. fistulosa might display a deeper purple color than the eastern varieties, this divergence has not been assessed at scale. We process over 40,000 community science photographs of M. fistulosa to identify flowers and extract color. We demonstrate that the flowers of the montane western variety have lower lightness and higher chroma, corresponding to a deeper violet color, than those of eastern M. fistulosa . Our approach and validation provides a scalable framework for phenotyping community science images and enables analysis of geographic color variation in other widespread species.

Why it matches plant phenotyping methods市民科学画像から花を同定し、花色を抽出・検証するスケーラブルな表現型解析パイプラインが研究の中心であり、植物器官の形質を直接推定している。

abstractWe process over 40,000 community science photographs of M. fistulosa to identify flowers and extract color.

Study on the germination rate of maize seeds based on improved YOLOv8n model. →

The germination potential of corn seeds, a key index for assessing their quality and directly associated with the ultimate corn yield, is currently defined in a way that cannot effectively portray the seed germination rate, and the prevalent measurement methods are traditional, consuming substantial process resources. To tackle these issues, this paper employs a public corn seed germination dataset, adds noise to it to simulate real - world production conditions, and ultimately acquires a dataset comprising 8148 images. It then proposes an enhanced YOLOv8 target detection model, EBS - YOLOv8, for detecting corn seed germination. Specifically, the ECA lightweight attention mechanism is introduced to decrease small - target feature loss, assist in accurate target recognition, and remove redundant features; simultaneously, the P2BiFPN multiscale feature fusion technique is utilized to boost the detection ability for small targets; furthermore, the ScConv convolution is adopted to enhance the feature - extraction capacity and improve detection accuracy. Combined with the improved model, this paper also proposed a mathematical modeling algorithmnew method for measuring seed germination potential and observing seed germination rate. The results indicate that the proposed model attains a mean average precision at 50% Intersection over Union (mAP50) value of 98.9%, a mean average precision in the range of 50% - 95% Intersection over Union (mAP50 - 95) value of 95.8%, an accuracy of 96.7%, and a recall of 96.3%. In comparison with the original model, the mAP50 has increased by 0.9% and the mAP50 - 95 value has witnessed a 3.7% increment. The experiments have demonstrated that the research method for germination potential put forward in this paper can effectively depict the rate variation of seeds during the germination process, thus offering a novel perspective for future research on seed germination potential.

Why it matches plant phenotyping methodsトウモロコシ種子の発芽状態・発芽率を画像から検出・定量する改良YOLOv8モデルと数理測定法を開発し、データセット上で性能検証しているため、植物フェノタイピング手法が中心である。

abstractIt then proposes an enhanced YOLOv8 target detection model, EBS - YOLOv8, for detecting corn seed germination.

An advanced deep learning method for pepper diseases and pests detection. →

Despite the significant progress in deep learning-based object detection, existing models struggle to perform optimally in complex agricultural environments. To address these challenges, this study introduces YOLO-Pepper, an enhanced model designed specifically for greenhouse pepper disease and pest detection, overcoming three key obstacles: small target recognition, multi-scale feature extraction under occlusion, and real-time processing demands. Built upon YOLOv10n, YOLO-Pepper incorporates four major innovations: (1) an Adaptive Multi-Scale Feature Extraction (AMSFE) module that improves feature capture through multi-branch convolutions; (2) a Dynamic Feature Pyramid Network (DFPN) enabling context-aware feature fusion; (3) a specialized Small Detection Head (SDH) tailored for minute targets; and (4) an Inner-CIoU loss function that enhances localization accuracy by 18% compared to standard CIoU. Evaluated on a diverse dataset of 8046 annotated images, YOLO-Pepper achieves state-of-the-art performance, with 94.26% mAP@0.5 at 115.26 FPS, marking an 11.88 percentage point improvement over YOLOv10n (82.38% mAP@0.5) while maintaining a lightweight structure (2.51 M parameters, 5.15 MB model size) optimized for edge deployment. Comparative experiments highlight YOLO-Pepper's superiority over nine benchmark models, particularly in detecting small and occluded targets. By addressing computational inefficiencies and refining small object detection capabilities, YOLO-Pepper provides robust technical support for intelligent agricultural monitoring systems, making it a highly effective tool for early disease detection and integrated pest management in commercial greenhouse operations.

Why it matches plant phenotyping methodsコショウの病害を画像から検出する深層学習手法を開発し、注釈画像データセットと複数モデルで性能比較・検証しており、植物の病害状態の取得方法が中心です。害虫検出も含みますが、病害検出の技術的貢献が明確なため採用します。

abstractthis study introduces YOLO-Pepper, an enhanced model designed specifically for greenhouse pepper disease and pest detection

DeepD&Cchl: an AI tool for automated 3D single-cell chloroplast detection, counting, and cell type clustering →

Chloroplast density in cells varies among different types of cells and plants. In current single-cell spatiotemporal analysis, the automatic detection and quantification of chloroplasts at the single-cell level is crucial. We developed DeepD&Cchl (Deep-learning-based Detecting-and-Counting-chloroplasts), an AI tool for single-cell chloroplast detection and cell-type clustering. It utilizes You-Only-Look-Once (YOLO), a real-time detection algorithm, for accurate and efficient performance. DeepD&Cchl has been proved to identify chloroplasts in plant cells across various imaging types, including light microscopy, electron microscopy, and fluorescence microscopy. Integrated with an Intersection Over Union (IOU) module, DeepD&Cchl precisely counts chloroplasts in single- or multi-layered images, while eliminating double-counting errors. Furthermore, when combined with Cellpose, a single-cell segmentation tool, DeepD&Cchl enhances its effectiveness at the single-cell level. By counting chloroplasts within individual cells, it supports cell-type-specific clustering based on chloroplast number versus cell size, offering valuable morphological insights for single-cell studies. In summary, DeepD&Cchl is a significant advancement in plant cell analysis. It offers accuracy and efficiency in chloroplast identification, counting and cell-type classification, providing a useful tool for plant research.

Why it matches plant phenotyping methods植物細胞画像から葉緑体を検出・計数し、細胞型をクラスタリングするAIツールの開発が中心であり、植物の形態的状態を定量化するフェノタイピング手法に該当する。

abstractWe developed DeepD&Cchl (Deep-learning-based Detecting-and-Counting-chloroplasts), an AI tool for single-cell chloroplast detection and cell-type clustering.

Disorder Severity Classification in Tomato Based on Weight Cluster Loss and Convolutional Neural Network →

Abstract Agricultural productivity is a crucial determinant of economic stability. Within the agricultural sector, particularly in tomato production, the impact of plant diseases and pests poses a significant challenge. Detecting the severity of disorders in tomato plants is essential for addressing these challenges. Achieving accurate and rapid detection is imperative for developing early treatment strategies, ultimately minimizing economic losses. While various researchers have explored solutions using convolutional neural network (CNN) models to identify and classify disease severity in tomatoes, the limited availability of training data has led to overfitting issues and inter-class similarity, resulting in suboptimal performance measures. To address the overfitting problem arising from insufficient data, this research proposes a deep transfer-based framework. Three CNN models i.e., AlexNet, SqueezeNet, and InceptionV3 are employed to classify disease severity in tomato plants, specifically targeting tomato late blight, tomato early blight, tomato leaf mold, tomato bacteria spot, and healthy tomato leaves using the PlantVillage dataset. The study incorporates a weighted-cluster loss function to mitigate inter-class similarities. Computational accuracy serves as the performance metric. Following experimentation, InceptionV3 demonstrated the highest classification accuracy at 93.66%, surpassing AlexNet (83.03%) and SqueezeNet (80.09%). Consequently, the proposed system functions as a decision support tool for farmers, aiding in the identification of disorder severity in tomato plant leaves.

Why it matches plant phenotyping methodsトマト葉の病害重症度という植物状態をCNNで画像分類し、重み付きクラスタ損失と複数モデル比較を中心的に評価しているため、植物フェノタイピング手法として採用。

abstractthis research proposes a deep transfer-based framework

Plant recognition of maize seedling stage in UAV remote sensing images based on H-RT-DETR. →

The real-time monitoring and counting of maize seed germination at seedling stage is of great significance for seed quality detection, field management and yield estimation. Traditional manual monitoring and counting is very time-consuming, cumbersome and error-prone. In order to quickly and accurately identify and count maize seedlings in a complex field environment, this study proposes an end-to-end maize seedling plant detection model H-RT-DETR (Hierarchical-Real-Time DEtection TRansformer) based on hierarchical feature extraction and RT-DETR (Real-Time DEtection TRansformer). H-RT-DETR uses Hierarchical Feature Representation and Efficient Self-Attention as the backbone network for feature extraction, thereby improving the network's ability to extract features of maize seedling stage in UAV remote sensing images. Through experiments on the UAV remote sensing data set of maize seedling stage, the mean Average Precision mAP0.5-0.95, mAP0.5 and mAP0.75 of the improved H-RT-DETR model reached 51.2%, 94.7% and 48.1%, respectively, and the Average Recall (AR) reached 68.5%. In order to verify the efficiency of the proposed method, H-RT-DETR is compared with the widely used and advanced target recognition methods. The results show that the detection accuracy of H-RT-DETR is better than that of the comparison methods. In terms of detection speed, the H-RT-DETR model does not require Non-Maximum Suppression (NMS) post-processing operations, the Frames Per Second (FPS) on the test dataset reaches 84f/s, which is 19,12,11 and 21 higher than that of YOLOv5, YOLOv7, YOLOv8 and YOLOX, respectively, under the same hardware environment. This model can provide technical support for real-time detection of maize seedlings under UAV remote sensing images in terms of both detection accuracy and speed (see https://github.com/wylSUGAR/H-RT-DETR for model implementation and results).

Why it matches plant phenotyping methodsUAV画像からトウモロコシ幼苗を検出・計数する深層学習手法を開発し、既存手法と精度・速度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ある。

abstractIn order to quickly and accurately identify and count maize seedlings in a complex field environment, this study proposes an end-to-end maize seedling plant detection model H-RT-DETR

Deep Learning VGG19 Model for Precise Plant Disease Detection →