Automated fruit counting and yield estimation systems are necessary for efficient orchard management. This study presents a computer vision system based on video multi-object-tracking for fruit load estimation in apple orchards and provides a comprehensive analysis of the system performance under diverse scanning conditions. The system integrates fruit detection, tracking, localization within orchard, and fruit load map generation. Experiments were carried out in an experimental apple orchard containing 420 apple trees. Data was collected with two different RGB-D sensors (Azure Kinect DK and ZED 2) at three different scanning distances (125 cm, 175 cm, and 225 cm) on two different dates prior to the harvest. Comparing the two evaluated sensors, Azure Kinect provided more consistent performance across different dates. Results also show that the longer scanning distance improves accuracy due to seeing the full tree view gives better fruit counts than close partial views. Between the two dates, best results were achieved near harvest due to fruit color at this stage, achieving a Mean Absolute Percentage Error (MAPE) of 6.91 % and a determination coefficient (R 2 ) of 0.733 (using ZED2 sensor at 225 cm distance). Finally, a test comparing scanning from one or both sides of the tree row showed that bilateral scanning improved fruit load estimation at the stretch level by incorporating information from both sides of the canopy. The results of this work demonstrate the effectiveness of the video fruit tracking systems as a useful tool for automating fruit load estimation.
Why it matches plant phenotyping methods動画ベースの果実検出・追跡手法を開発・評価し、リンゴ樹の果実負荷量を推定することが研究の中心であるため、植物フェノタイピング手法として含める。
abstractThis study presents a computer vision system based on video multi-object-tracking for fruit load estimation in apple orchards and provides a comprehensive analysis of the system performance under diverse scanning conditions.
High-throughput phenotyping of lettuce seedlings is highly prone to background confusion because the seedlings are small, have weak textural features, and exhibit spectral reflectance similar to that of the substrate. Traditional single-visual-modality approaches struggle to achieve reliable structural and physiological characterization simultaneously under the repetitive backgrounds and dense arrangements typical of greenhouse tray cultivation. To address these challenges, we establish a multimodal 3D phenotyping framework tailored for controlled agriculture environments, enabling the quantification of structural and physiological characteristics of lettuce seedlings. This framework is based on an unmanned ground vehicle (UGV) platform integrating a RGBD camera and a quad-band multispectral sensor which are rigidly coupled and synchronously triggered. An alignment module based on established feature matching algorithm is introduced to register the misalignment between source multispectral and RGBD images. Subsequently, we design a novel dual-backbone instance segmentation network, MS-SegNet, to enhance segmentation accuracy by hierarchically fusing geometric information with multispectral features. A robust 3D metric pose estimation pipeline, incorporating standard SfM initialization, scale recovery, and generalized ICP refinement, is constructed to generate 3D point clouds with spectral attributes and semantic labels. Finally, key structural and physiological phenotype parameters of each seedling are calculated based on the 3D semantic multispectral point clouds. Experiments demonstrate that MS-SegNet achieves significant advantages in instance segmentation of lettuce seedlings with mAP@50:95 = 0.854. The metric 3D pose estimation pipeline exhibits reliable performance under complex controlled conditions. The quality of the 3D reconstructions is indirectly validated through downstream structural trait extraction. The estimated seedling height and crown width show high correlation with manual measurements, achieving R 2 values of 0.8379 and 0.918, and RMSE values of 10.94 mm and 11.56 mm, respectively. Overall, by systematically integrating these adapted components with the novel segmentation architecture, this framework achieves stable performance improvements in 3D reconstruction, instance segmentation, and phenotypic analysis under greenhouse conditions. It provides a scalable, integrated technical solution for non-destructive, high-throughput phenotyping of crop seedlings in controlled environments.
Why it matches plant phenotyping methodsRGBD・マルチスペクトル・UGVを統合した3Dフェノタイピング基盤を開発し、分割・再構成・構造／生理形質抽出を検証しており、フェノタイピング手法が研究の中心である。
abstractwe establish a multimodal 3D phenotyping framework tailored for controlled agriculture environments, enabling the quantification of structural and physiological characteristics of lettuce seedlings.
Efficient facility-scale tomato ripeness monitoring remains difficult in greenhouses where uneven terrain limits conventional wheeled and rail-guided platforms and planar cameras provide restricted coverage. This study developed a wheel-legged quadruped monitoring system integrating LiDAR, a depth camera, and a panoramic camera. An adaptive gait-switching strategy supported navigation across heterogeneous terrain. Panoramic images were projected into six perspective views, and the left and right views were processed using a YOLOv8-based ripeness recognition model. Time-synchronized detections and robot poses were fused to map ripeness observations into three-dimensional greenhouse coordinates. Five field experiments in a commercial tomato facility demonstrated autonomous row traversal, inter-row transition, and avoidance of pedestrians, obstacles, and cultivation boundaries. The recognition pipeline continuously identified multiple ripeness stages under variable illumination, foliage occlusion, and robot motion, while the spatial fusion procedure produced a facility-scale three-dimensional ripeness distribution. The integration of terrain-adaptive quadruped mobility, panoramic perception, and spatial mapping provides a practical framework for continuous ripeness monitoring and can support targeted harvesting, yield forecasting, and crop management.
Why it matches plant phenotyping methodsトマト果実の成熟度を画像認識で取得するロボット型フェノタイピングシステムを開発し、実環境で評価しており、表現型取得法が研究の中心である。
abstractThis study developed a wheel-legged quadruped monitoring system integrating LiDAR, a depth camera, and a panoramic camera.
Accurate estimation of pasture biomass is essential for determining cattle stocking rates and grazing durations. The objective of this study was to comparatively evaluate five sensor-based systems for estimating aboveground Bermudagrass (Cynodon dactylon) biomass and identify the leading sensing approach for continued development and broader validation. The five systems included Structure-from-Motion (SfM), Ultrasound Sensor and Ski (US-Ski), Inertial Measurement Unit and Ski (IMU-Ski), Inertial Measurement Unit and Roller (IMU-Roller), and Depth Camera (DC). These systems were deployed on unmanned aerial and ground vehicles to measure crop height under identical field conditions. Regression models relating measured crop height to wet biomass yield (WBY) were developed as a common calibration framework for statistically comparing sensor performance. These empirical allometric equations were intended to support comparative benchmarking of the sensing systems and were not developed as final operational biomass prediction models for immediate field deployment. The influence of vegetation coverage on yield predictions generated by the crop height-based equations was also examined. The results indicated that the IMU-Ski system demonstrated the strongest overall comparative performance (R2 = 0.97; SeY = 1112 kg-wet/ha), followed by the DC system (R2 = 0.97; SeY = 1132 kg-wet/ha). Based on its overall benchmarking performance, including calibration accuracy, residual error and simplicity, the IMU-Ski system was identified as the leading sensing approach for continued development and broader validation among the five evaluated methods. The results also indicated that addition of vegetation coverage into the crop height-based regression models did not significantly improve prediction accuracy under the experimental conditions evaluated.
Why it matches plant phenotyping methods複数のセンサーシステムによる牧草バイオマス推定を比較・校正・ベンチマークしており、植物形質の取得法と技術性能の評価が研究の中心です。
abstractcomparatively evaluate five sensor-based systems for estimating aboveground Bermudagrass (Cynodon dactylon) biomass
Conventional methods of plant distance and volume measurements are limited by low efficiency, limited spatial coverage, and high measurement error. LiDAR and RGB-D imaging offer cost-effective, precise, and non-destructive techniques for plant distance and volume measurements. This study aimed to measure cabbage height, volume, and distance using LiDAR and RGB-D imaging. The sensors were mounted on a 1.6 kW electric field scouting platform (EFSP) for data collection. Point cloud (PCD) data were collected using LiDAR, whereas data processing, visualization, and measurements were done using commercial software and open-source programming scripts. A total of 20 cabbage plants were analyzed. LiDAR data processing included data frame screening, outlier removal, denoising, voxelization, and generation of 3D PCD density maps. Depth image processing included importing raw data and metadata shaping using intrinsic camera parameters, visualization, extraction of depth points, and pixel-level measurements of distances and volume. RGB image processing involved image conversion, segmentation, normalization, binary masking, mask cleaning, region extraction of cabbages, separation of ROI and preparation of contours, Delaunay triangulation and convex hull preparation, ROI overlay, bounding box preparation, sharing boundary between two boxes, conversion to pixel distances, and for visualization, plant height, volume measurements, and center to center distance measurement for measuring the plant distance. LiDAR demonstrated higher measurement accuracy for cabbage plant height, circumferential volume (geometric canopy volume), and plant distance, followed by RGB-D imaging, while RGB imagery showed comparatively lower performance under the study field conditions. Overall, LiDAR and RGB-D imaging provided reliable and non-destructive approaches for cabbage geometric characterization under field conditions, although accurately capturing complex plant geometry remains challenging. Positive and negative values of bias represent the over- and under-estimated results, respectively. Future studies should include larger and more diverse plant datasets exhibiting diversified size, shape, and geometric structure to further improve the robustness and general applicability of the proposed sensing approaches.
Why it matches plant phenotyping methodsLiDAR、RGB、RGB-Dを用いてキャベツの高さ・体積・株間距離を取得し、処理手順と測定精度を比較評価する手法研究であり、植物表現型取得が中心である。
abstractThis study aimed to measure cabbage height, volume, and distance using LiDAR and RGB-D imaging.
Abstract Depth cameras measure the distance to a fruit's surface, whereas converting its silhouette into physical dimensions requires the depth of its center; corrections for this offset have so far been empirical, and therefore bound to the crop, sensor, and dataset they were fitted on. This paper derives the correction analytically. For a spherical fruit, integrating the surface-depth distribution over the visible hemisphere yields a closed-form sphere correction whose coefficient follows from sampling geometry, together with a theoretical justification of the median mask depth as the representative statistic. Combined with deep instance segmentation on RGB-D imagery of hydroponic melons, the empirically optimal coefficient coincided with the derived value, and the pipeline reached R 2 of 0.966 for fruit length (MAE 1.43 mm), 0.959 for width (1.84 mm), and 0.861 for end-to-end fresh weight (MAPE 4.6%). The analytical form made the measurement transferable. Applied unchanged to cylindrical mini-cucumbers, the pipeline held mm-level accuracy (width MAE 0.52 mm; fresh weight R 2 0.955 after coe cient refitting), with the correction's negligibility predicted in advance by an R / Z corollary; across active-stereo and time-of-flight cameras, the optimal coefficients proved non-interchangeable, identifying the coefficient as a physical parameter that absorbs geometry, sensor physics, and fruit shape. A field system that fuses and cross-verifies the two sensors, with an error-propagation confidence gate and parameterized grading logic, reproduced 2-3% fresh-weight error and 90.9% confirmed-judgment grading accuracy over four validation sessions in a commercial greenhouse unseen during development. Throughout, geometric components transferred unchanged while learned and regression components required recalibration - a boundary the model predicts and the system itself monitors.
Why it matches plant phenotyping methodsRGB-D画像と深度補正を用いて果実の寸法・重量を推定する手法を開発し、異なる果形・センサー・圃場で精度検証しており、植物表現型取得が中心である。
abstractThis paper derives the correction analytically.
LiDAR (Light Detection and Ranging) and RGB-D camera imaging have emerged as essential tools in agricultural applications, particularly for plant size and distance measurements, enabling non-destructive, cost-effective, and precise estimation. The objective of this study was to measure the plant canopy dimensions and distance between apples using commercial LiDAR, and an RGB-D camera with a speed sprayer platform was used to determine whether LiDAR provides a higher measurement accuracy under field conditions. Data were collected in an apple orchard in Muju, Republic of Korea. Commercial 3D LiDAR, a terminal box, an RGB-D camera, a microcontroller, a power supply, and individual display monitors were integrated into a customized data acquisition (DAQ) box for LiDAR point cloud (PCD), RGB, and depth imagery data collection. Commercial software was used for data acquisition, data conversion (pcap to PCD), segmentation of regions of interest (ROI), and pre-processing of data. PCD processing and measurement consisted of data frame selection, data conversion, outlier removal, downsampling, denoising, ground point removal by filtering, voxelization, and density map generation using an open access programming language script. Depth image processing included importing raw data, shaping metadata using intrinsic camera parameters, visualizing depth images, extracting depth points, and measuring the plant canopy at the pixel level. RGB image analysis involved grayscale conversion, thresholding, segmentation of ROI, contour preparation, noise removal, and binary masking for eliminating the background. Estimated results were compared to measured results. LiDAR measurements showed the closest agreement with the measured results for plant height, canopy volume, plant spacing, and row distance, outperforming both RGB and depth imaging. Under field conditions, plant spacing and row distance were estimated with accuracies of 97.5% and 94.7%, respectively, exhibiting higher measurement accuracies than RGB and depth imagery data results. Despite some discrepancies due to complex plant geometry and dynamic data collection, the results support data collection strategies critical for precision horticulture.
Why it matches plant phenotyping methodsLiDARとRGB-D画像を用いたリンゴ樹の樹冠寸法・体積・樹間距離・列間距離の取得と精度比較が研究の中心であり、植物表現型計測手法の開発・検証に該当する。
abstractThe objective of this study was to measure the plant canopy dimensions and distance between apples using commercial LiDAR, and an RGB-D camera with a speed sprayer platform was used to determine whether LiDAR provides a higher measurement accuracy under field conditions.
To improve the instance segmentation accuracy of mango fruits and peduncles in complex mountainous orchard scenes and provide visual decision support for robotic harvesting, this study proposes a model-driven perception and picking-point localization method. Specifically, an RGB-D mango dataset was constructed under natural orchard conditions, covering strong illumination, shadows, backlighting, fruit overlap, branch and leaf occlusion, and peduncle crossing. An improved lightweight instance segmentation model named SHS-YOLOv8n-seg was then developed based on YOLOv8n-seg. StarNet_s1 was introduced as the backbone to enhance feature extraction under complex backgrounds. Furthermore, a high-frequency and spatial perception feature pyramid network was adopted to strengthen multi-scale feature fusion and improve the representation of slender peduncles. The SPPF module was used to expand the receptive field, and the parameter-free SimAM attention mechanism was introduced to enhance target responses while suppressing background interference. In the single-run comparison, the proposed model achieved Precision, Recall, mAP@50, and mAP@50:95 values of 89.62%, 88.17%, 90.19%, and 67.94%, respectively. Compared with the baseline YOLOv8n-seg model, these values increased by 2.42, 2.90, 2.75, and 2.57 percentage points, respectively. Moreover, fruit–peduncle matching, geometric constraints, RGB-D depth information, and PCA-based local direction estimation were combined to infer the picking point and recover its 3D coordinates from the segmentation masks. In the evaluation of 57 RGB-D images, the picking point position accuracy reached 98.2%, while the local peduncle direction accuracy reached 91.2%. Overall, the proposed method can accurately segment mango fruits and peduncles in complex natural environments and convert the segmentation results into picking point positions, 3D coordinates, and local direction information, thereby providing theoretical and technical support for intelligent mango harvesting robots.
Why it matches plant phenotyping methodsマンゴー果実・果梗の画像セグメンテーションと3D形状情報の抽出手法を開発・評価しており、単なる収穫対象の位置検出を超えて、再利用可能な植物器官の形態情報を取得する方法が中心である。
abstractfruit–peduncle matching, geometric constraints, RGB-D depth information, and PCA-based local direction estimation were combined to infer the picking point and recover its 3D coordinates from the segmentation masks
Precision agriculture increasingly relies on detailed structural information, such as canopy height and canopy volume, to enhance crop health monitoring and operational safety. However, existing methods based on costly LiDAR or RGB-D sensors are often impractical for large-scale deployment in dynamic and unstructured horticultural environments. Furthermore, conventional 2D segmentation and SLAM-based pipelines typically generate sparse, geometrically inconsistent semantic maps which are insufficient for actionable structural analysis in agricultural applications. To overcome these limitations, we propose a monocular 3D structural mapping framework tailored for horticultural plants via semantic scene completion. At inference, the proposed model takes a single RGB image as input and predicts voxel-wise geometry and semantics, from which task-oriented structural maps, including canopy height, canopy volume, and obstacle-aware traversability layers, are derived. Specifically, we first introduce a Depth-Aware Decoder Module that explicitly recovers depth in the spatial domain and fuses 2D-to-3D features, thereby mitigating depth ambiguity and reducing reliance on accurate pose. Second, an NCS-Guided Geometry Encoder is designed to inject normalized depth into voxel positional embeddings, enabling self-attention to perform global relational modeling within a depth-aware geometric coordinate system. In addition, a Global Encoder is utilized to refine local structural details, while an occupancy head produces the final 3D semantic completion outputs. We construct a horticultural 3D semantic scene dataset using an RGB-D sensor, which serves as a benchmark for evaluating our method, while the deployed model remains RGB-only. Extensive quantitative and qualitative experiments are conducted on both the Semantic-KITTI dataset and our dataset. On our dataset, the method achieves 82.31% occupancy IoU, 84.26% mIoU, and 86.25% precision. Beyond voxel-level evaluation, manual field measurements further show canopy height MAE values of 0.019-0.026 m and canopy volume proxy relative errors of 8.4%-11.4%. These results demonstrate the effectiveness of our approach in real-world agricultural scenarios, providing actionable structural insights for crop monitoring and autonomous robotic operations.
Why it matches plant phenotyping methods単眼RGB画像から植物の樹冠高・樹冠体積などの構造形質を推定する3Dフェノタイピング手法を開発し、データセット構築と実測検証も行っているため、方法が研究の中心である。
abstractwe propose a monocular 3D structural mapping framework tailored for horticultural plants via semantic scene completion.
Strawberry production in hydroponic systems requires precise and spatially explicit information on fruit load, maturity, and size to support harvest planning and quality control. This study presents an integrated computer vision and GNSS RTK framework that detects individual strawberries, estimates their maturity and physical dimensions, assigns fruits to operational size–maturity categories, and generates high-resolution yield maps in an open-field hydroponic system. RGB-D images are processed with YOLOv8-seg and a BoT-SORT-based tracking module to obtain instance-level masks, from which a color-based ripeness index and 3D point clouds are derived. After DBSCAN-based filtering, fruit length and width are computed in metric space and combined with ripeness percentage to assign each strawberry to one of nine operational size–maturity categories using predefined threshold rules. The system was evaluated in a commercial hydroponic crop in Arcabuco, Boyacá, Colombia, achieving accurate segmentation, with median IoU values up to 0.83 for bounding boxes and 0.71 for masks, and mean absolute errors of 2.51 mm in length and 1.85 mm in width with respect to Vernier caliper measurements. Yield maps aggregated in 1 m2 cells and by crop row revealed marked spatial variability in fruit density, maturity state, and size–maturity composition, allowing the identification of zones with a high concentration of fruits ready for harvest versus areas dominated by immature fruits. The proposed framework provides a practical and low-cost decision support tool for hydroponic strawberry management and may be adapted to other high-value horticultural crops after recalibration under different crop architectures, cultivars, and environmental conditions.
Why it matches plant phenotyping methodsRGB-D画像と追跡・3D解析により、イチゴ果実の成熟度、寸法、果実密度および収穫状態を個体レベルで推定し、精度検証も行う中心的なフェノタイピング手法研究である。
abstractThis study presents an integrated computer vision and GNSS RTK framework that detects individual strawberries, estimates their maturity and physical dimensions, assigns fruits to operational size–maturity categories, and generates high-resolution yield maps in an open-field hydroponic system.
Real-time strawberry volume estimation is a pivotal technology for automated harvesting and precision grading. However, conventional contact methods are prone to damaging fruits, while existing vision-based approaches struggle to balance high accuracy with low computational overhead. To address these challenges, this study proposes a two-stage real-time volume estimation framework coupling a red-green-blue-depth (RGB-D) sensor with an “Instance segmentation–Principal-axis slicing” framework. First, to precisely extract target contours in complex backgrounds, we designed Deformable Feature Aware-YOLO (DFA-YOLO) based on the YOLO11-seg architecture. This model enhances the geometric perception of irregular fruit edges and effectively overcomes the challenges of background noise and multi-scale variations, providing high-precision masks for subsequent spatial mapping. Subsequently, a principal-axis-slicing algorithm extracts the mask’s centroid and principal axis, perpendicularly slicing the mask into infinitesimal micro-slices. By computing and accumulating the pixel-space volume of these slices, the system converts them into precise 3D physical volumes based on RGB-D depth mapping. The entire system was deployed on an NVIDIA Jetson Orin edge computing platform and validated in a greenhouse. Experimental results demonstrate that the estimated volume highly agrees with the true volume, achieving a coefficient of determination (R2) of 0.945 and a mean absolute percentage error (MAPE) of 9.0%. Under typical operating conditions (1–5 targets per field of view), the system maintains an overall frame rate of 8–15 FPS, requiring only 55 ms for single-fruit estimation. This method exhibits favorable stability and lightweight efficiency under the tested greenhouse conditions, offering a reliable solution for real-time non-destructive crop phenotypic monitoring in computationally constrained agricultural environments.
Why it matches plant phenotyping methodsイチゴ果実の体積という植物器官形質を、RGB-D画像、インスタンスセグメンテーション、主軸スライシングで推定する方法を開発し、精度とリアルタイム性能を検証しているため、フェノタイピング手法が中心です。
abstractthis study proposes a two-stage real-time volume estimation framework coupling a red-green-blue-depth (RGB-D) sensor with an “Instance segmentation–Principal-axis slicing” framework.
Accurate yield estimation and crop load monitoring are essential for precision orchard management, supporting targeted fertilization, pruning, thinning, harvest planning, and marketing decisions. However, reliable in-situ monitoring remains challenging because commercial orchards are characterized by severe canopy occlusion, fruit overlap, heterogeneous tree architecture, variable illumination, and complex backgrounds. This review synthesizes advances in multi-modal sensing and deep learning for orchard yield estimation, breaking down the paradigm into intermediate fruit-counting or crop-load monitoring steps and supplementary spectral quality-assessment dimensions. First, yield-related indicators are summarized, including direct phenotypic traits such as fruit number, size, volume, and spatial distribution, as well as indirect structural and physiological proxies such as canopy volume, vegetation indices, flowering intensity, and spectral maturity attributes. Second, representative sensing devices and carrying platforms are reviewed, including red-green-blue (RGB) cameras, red-green-blue-depth (RGB-D) sensors, light detection and ranging (LiDAR), hyperspectral and multispectral systems, unmanned ground vehicles (UGVs), and unmanned aerial vehicles (UAVs). Third, the evolution of estimation methods is discussed, from traditional image processing and machine learning to object detection, instance segmentation, multi-object tracking, point-cloud analysis, remote-sensing regression, and multi-modal fusion. The review shows that no single sensor or algorithm can satisfy all orchard monitoring requirements. Ground-based vision and depth sensing are more suitable for fine-scale fruit counting and sizing, whereas UAV and spectral sensing provide advantages for regional yield mapping and quality-enhanced assessment. Future research should emphasize occlusion-aware perception, robust cross-environment generalization, lightweight edge deployment, standardized benchmarks, and integrated quantity-quality monitoring frameworks for actionable crop load management.
Why it matches plant phenotyping methods果実数・サイズ・体積・空間分布などの植物形質を対象に、センシング機器と画像解析・深層学習による収量推定法を体系的にレビューしており、フェノタイピング手法が中心である。
abstractThis review synthesizes advances in multi-modal sensing and deep learning for orchard yield estimation, breaking down the paradigm into intermediate fruit-counting or crop-load monitoring steps and supplementary spectral quality-assessment dimensions.
Accurate and non-destructive volume estimation of agricultural products is essential for precision agriculture, yet remains challenging when transitioning from controlled laboratory conditions to complex orchard environments. Although 2D image-based volume estimation methods provide a cost-effective and scalable solution, existing studies are fragmented and lack a unified perspective on their real-world applicability. This review presents a systematic synthesis of 2D image-based volume estimation methods, explicitly framed through the laboratory-to-orchard transition. We categorized existing volume estimation approaches according to the sensing modality into monocular RGB-based approaches and depth-assisted methods, and further reviewed them based on the image processing methods. A key finding is that high-precision geometric estimation can be achieved in laboratory environments, whereas deep learning and RGB-D fusion have driven a shift from conventional geometric modeling toward data-driven and hybrid learning frameworks in orchard settings. However, 2D image-based volume estimation remains fundamentally limited by scale ambiguity, severe occlusion, and sensitivity to illumination and background variability in real orchard environment. Overall, this review provides a unified perspective for understanding volume estimation methodology across environments and offers guidance for developing robust, scalable, and field-deployable volume estimation systems for real-world agricultural applications.
Why it matches plant phenotyping methods農産物の2D画像から体積を推定する手法を体系的にレビューしており、植物器官・果実の形態形質取得方法が中心である。
abstractThis review presents a systematic synthesis of 2D image-based volume estimation methods, explicitly framed through the laboratory-to-orchard transition.
Commercial greenhouse cucumber production is graded by fruit length, which drives harvest scheduling, labour allocation, and logistics. Manual measurement with thread or caliper is accurate but infeasible at commercial scale. This paper presents CucumberVision, a non-contact length estimation framework using an Intel RealSense D435 RGB-D camera. A YOLO26n instance segmentation model locates cucumbers, and SAM (ViT-B backbone) refines each detection to a pixel-precise mask. Five methods are evaluated under matched conditions: (M1) a dominant-axis skeleton scan-line baseline; (M2) PCA on the bounding-box depth point cloud; (M3) SAM mask with medial-axis skeletonisation; (M4) a hybrid keypoint-guided approach using a YOLO26-pose model predicting five anatomical landmarks (KP0--KP4) with piecewise 3D arc-length; and (M5) a novel medial arc spline method fitting a cubic spline through the 3D medial axis of the SAM mask and computing arc length by trapezoidal integration -- the first such application to elongated vegetable measurement. All methods share five-frame burst depth averaging, colour-stream intrinsic alignment, and adaptive method selection with cascading fallbacks ensuring 100% coverage. A benchmark of 48 captures across seven cucumbers in three size categories (small ~8 cm, medium ~13 cm, large ~25 cm) with thread-based ground truth establishes a significant accuracy hierarchy: M1 (MAPE 9.68%) > M2 (5.31%) > M4 (5.51%) > M3 (5.82%) > M5 (4.13%). M5 significantly outperforms all competitors at Bonferroni-corrected alpha=0.0125. A secondary contribution is identifying a 12--18% length underestimation caused by using depth-stream rather than colour-stream intrinsics after rs.align(rs.stream.color) -- an under-reported error source. The complete system is released open source and runs in real time on a single consumer-grade GPU.
Why it matches plant phenotyping methodsRGB-D画像からキュウリ果実長を推定する手法を開発し、複数手法との比較検証、実測値によるベンチマーク、誤差要因分析まで行っており、植物形質取得が研究の中心です。
abstractThis paper presents CucumberVision, a non-contact length estimation framework using an Intel RealSense D435 RGB-D camera.
Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · checked 13 Sept 2026
Modern agriculture operates at an unprecedented crossroads, it must simultaneously accelerate crop yields to feed an expanding global population and adapt to the severe, fluctuating pressures of climate change, structural soil degradation, abiotic water deficits, and evolving biological threats. Historically, selecting resilient crop varieties and implementing field-scale management strategies relied extensively on destructive, labor-intensive, and fundamentally subjective visual metrics. This manual processing approach has long been recognized as the primary operational bottleneck in agricultural advancement.To bridge the gap between rapidly expanding genomic data and actual field performance, the systematic, non-destructive quantification of structural and functional plant traits, plant phenotyping, has emerged as a transformative frontier. By integrating high-throughput engineering, multi-scale remote sensing, deep learning, and advanced molecular biology, modern phenotyping transitions crop science away from qualitative estimation toward highly reproducible, multidimensional data frameworks. This Research Topic presents new advances in advanced 3D reconstruction and deep semantic segmentation at the seedling stage; amodal fruit segmentation, morphological extraction, and early water-stress diagnostics; high-throughput in-field seedling counting and dynamic density modeling; multimodal foundation models, network pruning, and intelligent phytoprotection; aerial and spaceborne remote sensing for canopy analysis and weed monitoring; plant physiology, functional spectroscopy, and functional genomics under abiotic stress; and automated diagnostics for real-time orchard scouting and vineyard management.Automating the characterization of complex spatial layouts under controlled or greenhouse environments is essential for early variety selection and early-stage structural evaluation. Several contributions within this volume provide key breakthroughs in navigating overlapping tissues, severe occlusions, and low-contrast edge regions. showcases how substituting standard convolutions with deformable convolutions enables deep neural networks to accurately isolate the main stem of mature, high-density crops like soybeans. This architecture overcomes the traditional challenges of color mimicry and severe occlusion by pods and leaves, achieving an outstanding mIoU of 90.58% and providing reliable indices for lodging resistance and structural yield modeling (R 2 = 0.9746).Accurately extracting fruit morphology under commercial greenhouse conditions remains heavily constrained by overlapping crop structures, foliage cover, and variable shadows. Simple semantic masks typically fail when a target fruit is partially blocked, leading to a loss of key volumetric data.To resolve the challenge of hidden boundaries, Li, Yin, et al. (2025) developed CGA-ASNet, a specialized RGB-D amodal segmentation network driven by a Contextual and Global Attention (CGA) module designed to restore occluded tomato regions. Trained on a high-fidelity synthetic greenhouse dataset (Tomato-sim) generated via NVIDIA Isaac Sim's Replicator Composer and optimized with a mean coordinate fusion algorithm for real-world validation, this architecture expands the network's receptive field to predict the complete, hidden circular forms of occluded tomatoes, achieving an F@0.75 score of 94.2 and an amodal mIoU of 82.4%. This proves that simulation-to-real (Sim2Real) domain pathways can successfully decode full physical volumes under dense commercial canopies.Complementing this structural restoration, Yang, Li, et al. (2025) designed an integrated diagnostic framework to identify early water stress dynamics in greenhouse tomatoes. Built upon an optimized YOLOv11n core, their system integrates adaptive kernel convolutions (AKConv) into the network backbone's C3k2 modules and implements a recalibration feature pyramid detection head based on the specialized P2 small-target layer. This combination achieved a 5.4% increase in mAP50-95 for identifying fine phenotypic parts. By applying automated geometric analysis to the extracted bounding boxes, the system extracts plant heights and petiole count with low relative errors, feeding these phenotypic parameters into a Random Forest classification routine that flags water-stressed plants with 98% accuracy to guide targeted, automated drip irrigation.Accurate plant stands during early vegetative stages represent the foundational metric required to establish true field emergence rates, validate seed vigor across diverse breeding blocks, and perform early yield predictions.To solve the challenges of small targets, extreme spatial density, and adjacent leaf overlap, Zang et al. (2025) designed DM_IOC_fpn, a wheat seedling counting framework that balances local and global contextual features. By structuring a point-annotated dataset and embedding a densityenhanced encoder module, their network balances micro-scale spatial limits with macro-scale canopy structures. Optimized through a combined loss function tracking counting, classification, and regression parameters, this architecture achieved low error scores (RMSE = 2.91; MAE = 2.23), outperforming standard object-detection benchmarks in complex field environments.At the same time, scaling up to real-time aerial monitoring required major reductions in model complexity to support resource-constrained edge computers on autonomous aerial platforms. Feng, Nie, and Li (2025) engineered an ultra-lightweight YOLOv8n variant tailored for real-time maize seedling counting from high-speed UAV RGB overflights. By reparametrizing RepConv with HGNetV2, they constructed a lean Rep_HGNetV2 backbone, integrated a Bidirectional Feature Pyramid Network (BiFPN) for multi-scale feature alignment, and implemented a Task Dynamically Aligned Detection Head (TDADH). This architecture compressed total model parameters by 47% and reduced weight sizes to 3.5 MB while maintaining a 96.5% detection accuracy and an ultra-fast processing speed of 146.3 FPS, paving the way for low-cost, real-time field scouting.Automated phytoprotection requires machine-vision architectures capable of generalizing across highly diverse species, complex field conditions, and varying computational boundaries. A significant subset of the published papers addresses these challenges through foundation model adaptation, multi-modal alignment, and efficient network compression.A major paradigm shift presented in this collection involves moving away from task-specific training and toward foundation model adaptation. Chen, Ruan, et al. (2026) introduce a novel architecture integrating the DinoV3 foundation model with a Unet framework to achieve robust leaf lesion segmentation across diverse species (such as coffee and black gram). By incorporating a Spatial Prior Module (SPM), their approach surpassed standard benchmark networks by over 10.5% in IoU while reducing inference times by approximately 93.6%, demonstrating that highparameter foundation models can be highly optimized for resource-constrained edge devices in real-time scouting.To solve the perennial problem of limited training data for rare or emerging crop diseases, Cooper et al. ( 2026) developed an ingenious synthetic data generation pipeline. Combining 3D procedural leaf modeling in Blender with diffusion-based disease synthesis (Stable Diffusion fine-tuned with LoRA and ControlNet), they synthesized highly accurate plant disease images with perfect groundtruth annotation masks. When deployed in low-resource data settings, combining these synthetic pipelines with restricted real-world datasets consistently drives significant improvements in downstream segmentation tasks. To tackle specific, complex pathologies, Xu, Chang, et al. (2025) developed the TSSC deep learning model, which embeds three-neighbor channel attention paired with a complementary squeeze-and-excitation mechanism. This specific architecture minimizes structural degradation risks while pushing classification accuracy to 99.61% for highly complex pea leaf pathologies. Similarly, Feng, Liu, et al. (2025) tackled overlapping leaf occlusions and small lesion footprints in citrus groves with YOLO-Citrus, an optimized framework integrating C3K2-STA, ADown modules, and a Wise-Inner-MPDIoU loss function to strike a balance between edge computational constraints and field deployment.UAVs and high-resolution satellite imagery have expanded the operational scale of phenotyping from individual pots to vast breeding blocks and commercial fields, allowing researchers to capture macro-dynamic parameters over time.In complex canopy systems that defy standard top-down aerial sensing, such as single-staked white Guinea yams, Iseki et al. (2026) demonstrated the distinct advantage of utilizing multi-angle (combined nadir and oblique) UAV imaging configurations. When coupled with support vector regression, this method captures complementary canopy-structure information to model shoot biomass trajectories (R 2 = 0.79) across multiple years and management zones. These nondestructive, time-series datasets enabled the fitting of genotype-specific Richard's growth curves using Bayesian inference, isolating valuable genetic variations in early growth allocation.To capture full-season vertical physiological changes over large scales, Li, Yue, and Luo (2025) developed a hybrid CNN-LSTM-Attention (CLA) model designed to estimate the full-period Leaf Area Index (LAI) in rice using multi-temporal UAV multispectral imagery. By using the CNN layer to extract instantaneous spatial features, the LSTM block to process seasonal time-series intervals, and a self-attention mechanism to weight critical growth transitions, their platform achieved a high coefficient of determination (R 2 = 0.92) and kept relative root mean square errors (RRMSE) below 9%. This network minimized soil background noise during early vegetative stages (LAI values 1-
Why it matches plant phenotyping methods植物フェノタイピングの技術動向を扱うEditorialであり、画像解析、UAVセンシング、深層学習、形質抽出などの方法が中心的に整理されている。
Alternatives to soil-based horticulture, such as hydroponics, have been developed to respond to food distribution concerns for dense urban centers. A new system was developed to track an individual lettuce plant's growth in a hydroponic environment, utilizing streams of measured information and available models to continuously update the growth trajectory estimates for a plant. These "digital twin" models were integrated into an operating hydroponic greenhouse, with custom horticultural and sensor hardware to grow and measure relevant information. To aid in updating model parameters, plant yield was continuously measured with a custom neural network, using RGB-D images of the plants as an input. The network, trained on a collected dataset of 1300 images, was able to estimate mass within 1.5 g of the ground-truth value. After integration into the custom system, digital twin growth projections could approximate future yield between one and four days in the future, maintaining around a 2 g forecasting error.
Why it matches plant phenotyping methodsRGB-D画像から個体レタスの収量・質量を推定するニューラルネットワークと、センサー統合型の成長追跡基盤が研究の中心であり、植物形質の取得・予測手法を実質的に開発・検証している。
abstractA new system was developed to track an individual lettuce plant's growth in a hydroponic environment, utilizing streams of measured information and available models to continuously update the growth trajectory estimates for a plant.
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.
Reproduction assets foundThe paper explicitly states that links to its wheat spike dataset (multi-view images and 3D scans) and its analysis code are available via the authors' project webpage, which is an allowed URL. Other URLs (pyrender, CORDIS projects) are generic libraries or unrelated funding projects, not paper-specific assets.Dataset · publictance of around 2.5 m with a ground sampling distance of 0.3 mm (Fig. S1 a). The tagged and imaged spikes (Fig. S1 b) were sampled and ground truth volumes were acquired with a 3D light scanner (Shining 3D Einscan-SE V2, SHINING3D, Hangzhou, China) following the protocol of [ 76 ] . Links to the dataset and code can be found at https://oliviazum.github.io/3DKD-wheat/ . Detailed information about the dataset can be found in Sec. A .
3.3 Data Pre-Processing
Field images contained approximately 300-500 spikes per genotype within a plot of about 1.5 m 2 m^{2} . To reduce background inference, spike detection was first performed, and all subsequent processing was restricted to the detected regioOpen asset ↗lines:91-104Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · Europe PMC · checked 7 Sept 2026
Recent advancements in 3D reconstruction technologies have significantly transformed plant phenotyping, enabling precise, scalable, and automated trait extraction. Traditional manual phenotyping methods are increasingly being replaced by image-based approaches, such as photogrammetry, LiDAR, RGB-D sensing, and deep learning (DL)-based techniques. These tools allow for non-destructive, high-throughput measurements of plant morphology, structure, and physiological traits. This review synthesizes the state of the art in 3D reconstruction methods, including conventional geometric algorithms and emerging DL methods, and evaluates their application across diverse plant species. In addition, we discuss the sensing modalities, evaluation metrics, and crop-specific deployments. Although promising, current technologies still face challenges in terms of computational efficiency, scalability to outdoor environments, and generalizability across crop types. This review concludes by identifying research gaps and future directions for making real-time, field-deployable 3D phenotyping systems.
Why it matches plant phenotyping methods植物フェノタイピングにおける3D再構成技術と形質抽出を主題とする方法論レビューであり、評価指標やセンサー、応用を体系的に扱っている。
abstractThis review synthesizes the state of the art in 3D reconstruction methods
• Multisensor platform integrates RGB, depth, IR, and RTK-GPS data streams • Automated plant segmentation and 3D reconstruction extract plant traits in field conditions • System validation shows high correlation with manual and lab measurements • Public RGB-D lettuce dataset released to support reproducible AI research Accurate monitoring of leafy vegetable crops is essential to evaluate plant health, growth, yield, and quality, yet conventional methods based on manual measurements are labor-intensive and error-prone. This study proposes a data-driven framework for automated in-field monitoring of a lettuce crop based on multidimensional data acquired by a ground platform under various field conditions. Specifically, an advanced perception system is developed, including imaging and localization sensors to capture high-resolution visual, structural, and georeferenced information on the crop. An image processing pipeline is then proposed using zero-shot learning for plant segmentation, followed by 3D phenotyping techniques based upon computational geometry to automatically estimate plant biophysical traits, thus minimizing human input. An experimental trial conducted in a test field in Bari, Italy, between April and May 2025 validated the approach against manual and laboratory estimations. The results demonstrate strong correspondence between automated and reference measurements with a Pearson correlation coefficient r > 0.9 for key traits, confirming the potential of the framework. The influence of different nitrogen levels on the growing cycle is also evaluated, showing that the proposed system may provide a useful tool for decision support in lettuce crop monitoring and management.
Why it matches plant phenotyping methodsRGB・深度・IR等を統合したセンシング、植物セグメンテーション、3D形状解析による形質推定を開発し、手測定・実験室測定で検証しているため、フェノタイピング手法が中心である。
abstractAn image processing pipeline is then proposed using zero-shot learning for plant segmentation, followed by 3D phenotyping techniques based upon computational geometry to automatically estimate plant biophysical traits
• Depth informed trunk detection performed reliably under field conditions. • Trunk diameter estimates aligned closely with ground truth. • Transmission-based digital odometer provided reliable along-row positioning. • Full-block mapping demonstrated large-scale applicability. Trunk diameter is an important structural trait used to assess tree size, vigor, and long-term growth in orchard systems, but it remains difficult to measure efficiently at orchard scale. This study developed and validated a low-cost mobile imaging system for automated trunk diameter estimation and spatial mapping in tart cherry (Prunus cerasus) orchards. The system integrated RGB-D cameras, deep learning-based trunk detection and segmentation, three-dimensional depth reconstruction, and driveline-based digital odometry for spatial positioning under canopy conditions where GNSS performance was unreliable. Trunks were detected using a YOLO-based model, segmented using a Segment Anything Model (SAM)-based approach, and reconstructed in 3D from depth data to estimate real-world trunk diameter at 40 cm above the trunk base. The system was evaluated in a 15-year-old, 1 ha experimental orchard in Utah, USA, and then applied in a 14-year-old, 9.5 ha commercial orchard. Under unobstructed viewing conditions, trunk diameter estimates showed strong agreement with manual measurements, with a mean absolute error (MAE) of 0.95 cm, a root mean square error (RMSE) of 1.16 cm, and R 2 of 0.79. Across all 462 trees, automated per-tree averaging produced an MAE of 1.40 cm. In the commercial orchard, mapped trunk diameter patterns aligned with UAV-derived canopy height, reflecting underlying zones of tree vigor. These results show that mobile RGB-D imaging combined with driveline-based odometry can provide practical, cost-effective orchard-scale trunk diameter mapping under commercial field conditions.
Why it matches plant phenotyping methodsRGB-D画像、深度再構成、物体検出・セグメンテーション、オドメトリを統合し、樹幹径という植物構造形質を自動推定・空間マッピングする手法を開発・検証しているため。
abstractThis study developed and validated a low-cost mobile imaging system for automated trunk diameter estimation and spatial mapping in tart cherry (Prunus cerasus) orchards.
ABSTRACT This article describes a method for measuring the diameter of fruits with a near spherical shape using an RGB‐D camera in order to investigate fruit enlargement at different times of the year. In general, depth‐based measurement methods, when considering the diameter of an object with a near‐spherical shape, the diameter can be calculated by obtaining the Euclidean distance from the coordinates of the object's sides or by using the depth at the center of the object and the size of the object on the RGB image. However, it is difficult to accurately determine the diameter of an object because of errors in the calculated results due to the perspective projection of a general camera. Therefore, this study proposes a method to measure the diameter of fruits that have a shape similar to a sphere. Although this study focuses on young apple fruits, the proposed method can be applied to other agricultural crops, as well as to objects that are similar to spheres. In addition, we have also studied a correction that takes into account the rotation of the object so that the method can be applied to objects with circular cross‐sections.
Why it matches plant phenotyping methodsRGB-Dカメラを用いて果実径を測定する手法を提案・補正しており、植物形質の取得方法が研究の中心である。
abstractThis article describes a method for measuring the diameter of fruits with a near spherical shape using an RGB‐D camera
Plant phenotyping relevance match · UnverifiedCrossref · OpenAlex · Europe PMC · checked 15 Sept 2026
Computer vision and Internet of Things (IoT) technologies offer robust solutions for plant phenotyping, but traditional mainstream segmentation methods often fail in high-density plantings with overlapping foliage. This study introduces an integrated phenotyping system combining automated data capture and high-temporal RGB-D imaging using off-the-shelf hardware (Intel RealSense D435 and Raspberry Pi) to generate 3D point clouds of lettuce under controlled greenhouse conditions. While recent agricultural applications have shown limited success and required domain-specific adaptations, Segment Anything Model (SAM) and FastSAM were demonstrated to achieve exceptional zero-shot segmentation performance for individual lettuce plants in high-density arrangements without additional training. This capability effectively addresses the traditional challenges of species-specific parameter tuning and extensive training data requirements and fine-tuning. By mapping 2D segmentation masks to corresponding 3D point clouds, the system accurately extracted key phenotypic traits, namely plant height, length, and width, from which area and volume were subsequently estimated, showing strong correlations with manual measurements for Rex and Rouxai lettuce cultivars. This high-temporal, non-destructive monitoring provided unique insights into plant growth dynamics. The study highlights distinct growth patterns among these cultivars, underscoring the importance of tailored phenotyping approaches to optimise crop management strategies. By addressing the limitations of existing phenotyping methods, this work advances precision agriculture technologies, offering a cost-effective and efficient solution for monitoring dynamic crop growth with potential applications across various crops and growing conditions.
Why it matches plant phenotyping methodsRGB-D撮像、3D点群、セグメンテーションを統合した植物表現型取得システムを開発し、草丈・長さ・幅などを抽出して手測定と検証しているため、方法が中心的である。
abstractThis study introduces an integrated phenotyping system combining automated data capture and high-temporal RGB-D imaging using off-the-shelf hardware (Intel RealSense D435 and Raspberry Pi) to generate 3D point clouds of lettuce under controlled greenhouse conditions.
Computer vision and Internet of Things (IoT) technologies offer robust solutions for plant phenotyping, but traditional mainstream segmentation methods often fail in high-density plantings with overlapping foliage. This study introduces an integrated phenotyping system combining automated data capture and high-temporal RGB-D imaging using off-the-shelf hardware (Intel RealSense D435 and Raspberry Pi) to generate 3D point clouds of lettuce under controlled greenhouse conditions. While recent agricultural applications have shown limited success and required domain-specific adaptations, Segment Anything Model (SAM) and FastSAM were demonstrated to achieve exceptional zero-shot segmentation performance for individual lettuce plants in high-density arrangements without additional training. This capability effectively addresses the traditional challenges of species-specific parameter tuning and extensive training data requirements and fine-tuning. By mapping 2D segmentation masks to corresponding 3D point clouds, the system accurately extracted key phenotypic traits, namely plant height, length, and width, from which area and volume were subsequently estimated, showing strong correlations with manual measurements for Rex and Rouxai lettuce cultivars. This high-temporal, non-destructive monitoring provided unique insights into plant growth dynamics. The study highlights distinct growth patterns among these cultivars, underscoring the importance of tailored phenotyping approaches to optimise crop management strategies. By addressing the limitations of existing phenotyping methods, this work advances precision agriculture technologies, offering a cost-effective and efficient solution for monitoring dynamic crop growth with potential applications across various crops and growing conditions.
Why it matches plant phenotyping methodsRGB-D画像、3D点群、セグメンテーションを統合して植物形質を抽出するフェノタイピングシステムの開発・評価が中心であり、手動測定との相関検証も行っているため。
abstractThis study introduces an integrated phenotyping system combining automated data capture and high-temporal RGB-D imaging using off-the-shelf hardware (Intel RealSense D435 and Raspberry Pi) to generate 3D point clouds of lettuce under controlled greenhouse conditions.
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.
Reproduction assets foundThe paper's Data availability statement points to a public Zenodo deposit containing the datasets generated and analysed in this study (DMA exudation and Zn uptake measurements in rice).Dataset · publicthe experiments, developed the
525
methods and analysed the results. The experimental data were collected by C.R. assisted by
526
G.L.M., C.T. and D.J.W. Data analysis and writing of paper by all authors.
527
528
Data availability
529
The data sets generated and/or analysed during the current study are available on Zenodo,
530
https://zenodo.org/uploads/18184803
531
532
533
.
CC-BY 4.0 International license
perpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for this
this version posted March 18, 2026.
;
https://doi.org/10.64898/2026.03.16.71158Open asset ↗Zenodo · 18184803pdf-raw-page:21 lines:1-47Code / dataset availability confirmedOpenAlex · Europe PMC · checked 5 Sept 2026
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.
Reproduction assets foundThe paper's authors explicitly state their analysis code is publicly available on GitHub. Phenotype datasets (RGB-D orchard image sequences, LiDAR point clouds, manual trait measurements) are only available upon request, so they do not qualify as public assets.Code · publicOur code are available at https://github.com/zlhzau/InspectGaussian.git .Open asset ↗zlhzau/InspectGaussianlines:489-515Plant phenotyping relevance match · UnverifiedOpenAlex · checked 15 Sept 2026
• Modular RGB–D pipeline for real-time multi-attribute grading of strawberries. • ResNet-101 multi-head classifier predicted maturity, disease, and deformity jointly. • CycleGAN restored strawberry shape and texture from partially occluded views. • QH and MADM strategies converted multi-attribute predictions to grading decisions. • Field tests validated end-to-end perception-to-placement pipeline in farms. Strawberry grading was critical for meeting market standards, yet manual post-harvest sorting remained labor-intensive, damage-prone, and inconsistent, while most vision-based methods targeted single attributes and lacked integrated in-field operation. We presented a real-time RGB–D grading framework for tabletop-cultivated strawberries in greenhouse environments. The system combined instance segmentation (YOLOv11–Seg), occlusion-aware completion via CycleGAN, and a three-branch multi-task network based on ResNet-101 to simultaneously infer maturity, disease status (including asymptomatic fruit), and deformity. Pose normalization and depth alignment supported calibrated volume-to-mass regression for non-destructive weight estimation. To improve robustness to rare morphology, 500 deformed-fruit samples were synthesized with a Gemini-based generative model to mitigate class imbalance. Built upon this perception stack, we implemented two grading strategies: a deterministic Quality Hierarchy (QH) for real-time robotic routing and a weighted multi-attribute decision-making (MADM) scheme for flexible batch evaluation. The multi-attribute classifier achieved 93.65% overall accuracy (disease 94.09%, maturity 94.07%, shape 94.42%) with an inference latency of 33.65 ms (29.7 FPS). Depth-assisted mass estimation yielded average errors of 8.11% for complete fruits and 10.47% under occlusion after completion; in field deployment on marketable fruits routed to size grading, it achieved MAE/RMSE of 1.85/2.20 g with R 2 = 0.9384 (MAPE 8.48%) and a three-bin size-grade accuracy of 90.91% (100/110). In single-target harvesting mode on an RTX 4060 GPU (640x480, 30 fps), end-to-end latency was 87.56 ms per harvested target in complete mode and 197.56 ms when completion was invoked (CycleGAN 110.00 ms/instance). These results demonstrated a practical, non-destructive, occlusion-aware, and deployable multi-attribute grading solution for intelligent strawberry harvesting in real greenhouse scenarios.
Why it matches plant phenotyping methodsRGB-D画像解析でイチゴの成熟度、病徴、形状、重量を推定し、性能検証とロボット実装まで行う中心的なフェノタイピング手法研究。
abstractModular RGB–D pipeline for real-time multi-attribute grading of strawberries.
Controlled-environment agriculture (CEA) and circular production systems require coordinated monitoring of biological and physicochemical processes across trophic levels. This project report presents the implementation of a multi-trophic controlled-environment agriculture demonstrator that integrates computer-vision-based monitoring with established sensor infrastructure for aquaculture, poultry, plants, microalgae, duckweed, and insect modules. Stereo imaging and RGB-D systems are deployed for non-invasive quantification of fish biomass and plant growth, while continuous water-quality and environmental measurements (e.g., pH, dissolved oxygen, nitrate, ammonium, temperature, CO2) provide complementary process data. These data streams are synchronized within a shared database architecture to enable cross-module evaluation of nutrient dynamics, growth progression, and operational stability under real facility conditions. The implemented framework demonstrates how computer vision can extend conventional sensor-based monitoring by directly capturing biological performance indicators across aquatic, terrestrial, and microbial domains. While advanced predictive modeling and full digital twin simulation remain future development steps, the realized data-integration architecture establishes a structural foundation for the systematic evaluation of circular indoor food-production systems. The demonstrator illustrates how multimodal monitoring can support nutrient recirculation, transparency of biological variability, and data-driven assessment within controlled multi-trophic environments.
Why it matches plant phenotyping methodsステレオ画像およびRGB-Dによる植物生長の非破壊定量と、データ統合型モニタリング基盤の実装が中心的に記述されており、植物フェノタイピング基盤として収載対象です。
abstractStereo imaging and RGB-D systems are deployed for non-invasive quantification of fish biomass and plant growth
Field / plotLaboratory / benchtopNeRF / 3D Gaussian SplattingPhotogrammetry / SfM / MVSLiDAR / point cloudRGB-D / ToFWhole plant / canopy / plot / field2D/3D reconstruction
Three-dimensional point cloud (3DPC) data capture detailed geometric and structural plant traits beyond the capability of 2D imaging. When combined with artificial intelligence (AI), it offers a powerful, non-invasive tool for plant phenotyping, which is crucial for driving advancements in plant breeding and agriculture. However, challenges related to data complexity, limited datasets, and model generalization hinder 3DPC’s widespread adoption. To provide a comprehensive overview and guide future research in this area, we conducted a systematic literature review (SLR) following Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines by analysing 381 papers published between January 2017 and October 2025 from major databases. Our review examines the advantages, current status, limitations, and future directions of AI applications in 3DPC-based plant phenotyping. Our findings indicate a rapid increase in publications since 2022, with deep learning (DL) methods, especially pointwise MLP-based networks, driving much of this growth, with a notable recent surge in Transformer-based, Graph-based, and particularly Hybrid models that combine their strengths. Furthermore, novel methods like Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) are emerging as powerful tools for 3D reconstruction and scene synthesis. Time-of-Flight (ToF) and Structure from Motion and Multi-View Stereo (SfM-MVS) technologies remain the predominant 3DPC data acquisition techniques. Research in this area focuses on trees/shrubs and cereals, typically involving single-species studies. Although the overall use of public datasets remains low (18.9%), their adoption has significantly increased since 2020. Key limitations identified include: (1) a lack of standardized data collection and formats, (2) insufficient model robustness and generalization, especially from lab to field, (3) high computational demands, and (4) a reliance on species-specific models. The future of AI-driven 3DPC phenotyping hinges on overcoming these bottlenecks. Priority should be given to: developing field-deployable, computationally efficient models; exploring the potential of the foundation model; establishing diverse and standardized public datasets; and strengthening the integration of 3D phenomics with genomics to bridge the genotype-to-phenotype gap. This review provides a foundational roadmap to guide research in plant phenomics, crop breeding, and plant science.
Why it matches plant phenotyping methods3D点群とAIによる植物形質取得・解析を中心に扱う体系的レビューであり、植物フェノタイピング手法のレビューとして明確に適格。
titleAI-driven 3D point cloud analysis in plant phenotyping: A Systematic Review
Plant phenotyping relevance match · UnverifiedCrossref · OpenAlex · Europe PMC · checked 15 Sept 2026
Reconstruction of crop three-dimensional (3D) point clouds is essential for monitoring phenotypic parameters, like plant height and leaf area index (LAI), which is a critical phenotype predictor for smart crop breeding. The main 3D reconstruction technologies include image-based approaches, laser scanning, and depth camera methods. Among these methods, image-based structure-from-motion (SfM) is widely used due to its low cost and high accuracy. However, field crop canopy image data for high-resolution point cloud construction are often large-scale, unordered, and uncalibrated. Conventional SfM methods struggle with 3D reconstruction due to high computational costs and long processing times, delaying phenotypic analysis. To address this issue, we developed an improved global SfM algorithm, which increases the point cloud reconstruction speed by an average of 1.39 times compared to traditional incremental SfM methods and by more than 10 % on average compared to two mainstream global SfM algorithms. In addition, we integrated three types of predictors, point cloud features, color indices and texture features, through multi-feature data fusion and machine learning. A random forest algorithm for the prediction of LAI for a combined data set of four different crops, and using all three categories of predictors, achieved higher monitoring accuracy compared to using a single feature category (R²=0.78 vs R²=0.71–0.74). This new method, which includes an improved global SfM algorithm and a three-predictor fusion-based LAI monitoring approach, offers an efficient and reliable solution for precise crop phenotyping and continuous growth monitoring in complex field environments, enabling accurate assessment of crop morphology and developmental dynamics.
Why it matches plant phenotyping methods改良したSfMによる3D再構成と、特徴量融合・機械学習によるLAI推定を開発・評価しており、植物表現型取得手法が研究の中心である。
abstractwe developed an improved global SfM algorithm
This paper describes a method for measuring the diameter of fruits with a near spherical shape using an RGB-D camera in order to investigate fruit enlargement at different times of the year. In general depth-based measurement methods, when considering the diameter of an object with a near-spherical shape, the diameter can be calculated by obtaining the Euclidean distance from the coordinates of the object's sides or by using the depth at the center of the object and the size of the object on the RGB image. However, it is difficult to accurately determine the diameter of an object because of errors in the calculated results due to the perspective projection of a general camera. Therefore, this study proposes a method to measure the diameter of fruits that have a shape similar to a sphere. Although this study focuses on young apple fruits, the proposed method can be applied to other agricultural crops, as well as to objects that are similar to spheres. In addition, we have also studied a correction that takes into account the rotation of the object so that the method can be applied to objects with circular cross-sections.
Why it matches plant phenotyping methodsRGB-D画像から果実径を推定する手法の開発が研究の中心であり、植物器官の形態形質を直接測定するため。
abstractThis paper describes a method for measuring the diameter of fruits with a near spherical shape using an RGB-D camera
Identifying nutritional deficiencies at an early stage is crucial for maximizing yield production and ensuring healthy plants. Conventional methods generally rely on time-consuming analysis conducted by agronomic experts. To address this challenge, this study presents a data-driven approach for the early identification of nutritional deficiencies in hazelnut orchards. Different custom datasets, composed of images acquired in a real hazelnut orchard as well as in a controlled laboratory environment, are collected, and the performance of five state-of-the-art machine learning models in early detecting nutritional deficiencies is compared. In particular, ResNet, DenseNet, MobileNet, EfficientNet, and ConvNext models, along with a baseline based on support vector machines, are considered. Data augmentation techniques are introduced to synthetically increase the datasets, and their effectiveness is extensively evaluated. Additionally, a pipeline is designed to carry out the early identification of nutritional deficiencies onboard an agricultural robot. Experimental results on the early identification show that ConvNext achieves the highest performance: 81.79% accuracy and 0.8168 F1 score on a real-world dataset with four classes, and 75.54% accuracy with 0.7552 F1 score for the more challenging six-class scenario. Furthermore, the effectiveness of the integrated system is validated in preliminary laboratory experiments using a Turtlebot2 mobile base and a Franka Research 3 arm, equipped with RGB-D cameras. • Data-driven pipeline detects hazelnut nutrient deficiencies from leaf images. • Real and lab-acquired hazelnut leaf image datasets collected and publicly released. • ConvNext achieves 85.50% accuracy on 4-class hazelnut deficiencies in lab conditions. • 75.54% accuracy on 6-class real orchard dataset validates field robustness. • Two-stage pipeline with leaf detection and classification enables onboard robot monitoring.
Why it matches plant phenotyping methods葉画像から植物の栄養欠乏状態を推定する画像解析・機械学習パイプラインと、ロボット搭載システム、データセットを中心的に開発・評価しているため。
abstractthis study presents a data-driven approach for the early identification of nutritional deficiencies in hazelnut orchards
This data article describes a curated RGB-Depth image dataset captured using an Intel RealSense D435 stereo depth camera mounted on an autonomous mobile platform during field deployments at commercial baby broccoli farms in Victoria, Australia. The dataset comprises 1759 paired RGB images (640 × 480 pixels) and corresponding 16-bit depth frames acquired under both daytime (natural sunlight) and night-time (LED illumination) conditions, designed to support research in agricultural computer vision and robotic harvesting. Images were selected from 39,765 raw acquisitions through a reproducible Python curation pipeline applying quality filtering (blur detection, brightness thresholds, corruption detection), perceptual hash-based duplicate removal, and manual review. The final dataset includes 924 daytime and 835 night-time image pairs containing baby broccoli plants at various growth stages. The dataset provides RGB camera intrinsic parameters and pixel-aligned depth maps to enable 3D point cloud reconstruction. Potential applications include developing deep learning models for crop detection and segmentation, validating depth-based size estimation methods, and benchmarking illumination-robust vision systems. All data and curation code are publicly available under a CC BY 4.0 license.
Why it matches plant phenotyping methodsRGB-Depth画像データセットの構築と再現可能なキュレーションを中心とし、作物検出に加えてサイズ推定という植物形質の評価・ベンチマークに利用できるため。
titleA field-acquired RGB-Depth image dataset for computer vision-based baby broccoli detection and size estimation under varying illumination conditions.
Reproduction assets foundThe paper is a data article describing a public Mendeley Data repository containing the authors' field-acquired RGB-D baby broccoli image dataset (1759 image pairs, ground truth diameter annotations, camera intrinsics, and curation/annotation code), directly reproducing the paper's phenotyping measurements and analysisDataset · publicRepository name: Mendeley Data
Data identification number: 10.17632/px5p6zdk6k.3
Direct URL to data: https://data.mendeley.com/datasets/px5p6zdk6k/3Open asset ↗Mendeley Data · 10.17632/px5p6zdk6k.3html-lines:95-155Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · OpenAlex · checked 15 Sept 2026
Hyperspectral imaging (HSI) has emerged as a powerful tool for precision agriculture, enabling the non-destructive monitoring of crop biochemical and physiological traits. However, HSI alone lacks structural context, which limits its ability to accurately capture complex canopy architectures and organ-level traits. Integrating HSI with depth-sensing modalities such as Light Detection and Ranging (LiDAR), Red, Green, Blue, and Depth (RGB-D) cameras, and computational reconstruction technique such as photogrammetry enables the generation of three-dimensional hyperspectral point clouds, combining spectral richness with geometric fidelity. This multi-modal fusion enhances crop trait estimation, including biomass, leaf chlorophyll content, canopy height, leaf area, and stress indicators, while improving the robustness of phenotyping under occlusions, shadows, and varying illumination. Dimensionality reduction, feature selection, and machine learning approaches, including deep learning and explainable AI, are useful for handling high-dimensional hyperspectral data and extracting actionable agronomic insights. Moreover, the integration of thermal, radar, and Global Navigation Satellite System (GNSS) data further expands the capabilities of multi-modal sensing, enabling continuous, all-weather crop monitoring and accurate spatial referencing. Despite these advances, most studies to date focus on controlled environments, highlighting the need for field-based validation to ensure the reliability and scalability of HSI-depth fusion techniques. This review consolidates current knowledge on multi-modal hyperspectral and 3D crop reconstruction, highlighting methods, applications, and challenges, and outlines future directions for implementing high-throughput, real-time phenotyping and precision agriculture solutions.
Why it matches plant phenotyping methods植物形質推定のためのハイパースペクトル・深度センシング融合と3D再構成を中心に扱うレビューであり、フェノタイピング手法の方法論的整理が主題。
abstractThis review consolidates current knowledge on multi-modal hyperspectral and 3D crop reconstruction, highlighting methods, applications, and challenges
• ToF camera was evaluated for 3D phenotyping and canopy volume estimation. • Strong correlation was observed between ToF imagery and manual methods. • ToF technology could accurately estimate canopy volume in horticultural crops. • Study highlights the potential of ToF imagery for smart farm management. Noninvasive and accurate quantification of plant architectural traits remains a persistent challenge in horticultural crop breeding and high-throughput phenotyping (HTPP), largely due to labor constraints and limited availability of scalable 3D sensing technologies. Time-of-Flight (ToF) imaging offers a promising active sensing approach capable of generating high-resolution digital replicas of plant canopies for structural trait analysis. This study evaluated the feasibility of deploying 3D ToF imaging for digital phenotyping of three horticultural and ornamental crops, guava, brinjal, and jasmine with a focus on canopy volume estimation. High-resolution 3D point clouds were acquired for five plants per crop across three growth stages during the summer season of 2024. Structural parameters including plant height, width North–South (W NS ), width East–West (W EW ), and 3D aspect ratio were extracted and used to compute canopy volume via a voxel grid method. To validate accuracy, ToF-derived canopy volumes were compared against manually measured volumes estimated using the prolate spheroid volume (PSV) method. The ToF-based digital phenotyping framework demonstrated strong agreement with manual measurements across all crops. Guava exhibited the highest accuracy at the intermediate growth stage (R² = 1.0; RMSE = 0.0004), while brinjal (R² = 0.908; RMSE = 0.004) and jasmine (R² = 0.935; RMSE = 0.4351) showed robust performance at the full-grown stage. Those regression relationships were statistically significant ( p < 0.05), confirming the reliability of ToF-based canopy reconstruction for structural phenotyping. Overall, the findings demonstrate that ToF imaging provides a robust, noninvasive, and scalable framework for digital canopy phenotyping in horticultural crops, supporting precision breeding and structural trait monitoring applications in field environments.
Why it matches plant phenotyping methodsToFによる3D画像取得とキャノピー体積・構造形質の抽出を開発・評価し、手動測定との比較検証を行っており、植物表現型取得法が研究の中心である。
abstractToF camera was evaluated for 3D phenotyping and canopy volume estimation.
AppleField / plotLiDAR / point cloudRGB-D / ToFFruitObject detection2D/3D reconstructionSegmentation
One of the key challenges in orchard robots is accurately localizing occluded fruits in complex environments, especially when the fruit targets are split into multiple isolated regions within images. Traditional single-task network models exhibit limited capability in discerning fragmented targets that belong to the same fruit but are segmented into multiple spatially isolated regions within images. In addition, fruit localization largely relies on high-cost sensors or additional 3-D localization algorithms. To address this issue, we propose a fruit detection and centroid localization method based on a Multi-Task Wavelet-Enhanced YOLO (MT-WavYOLO) to enhance the success rate of robotic operations on occluded fruit targets. Initially, a lightweight semantic segmentation branch was integrated into the YOLOv8 backbone network to precisely segment exposed fruits, while retaining the original object detection branch to fully identify occluded fruits. To address the diminished sensitivity of conventional models to geometric profiles of heavily occluded fruits, a novel feature fusion module, C2f_WTConv, was designed by incorporating wavelet transform convolution, leveraging the multi-frequency robustness of wavelet representations to enhance the model’s feature extraction capabilities under complex orchard occlusions. Subsequently, a 3D frustum-based point cloud processing method was proposed, combining the detection results from MT-WavYOLO with the semantic segmentation masks to accurately localize occluded fruits. MT-WavYOLO demonstrated a 2%, 1.5%, and 2.2% improvement in Precision, Recall, and mAP50, respectively, on our custom-built dataset compared to the latest YOLOv10s model. Semantic segmentation performance, measured by Intersection over Union (IoU) and Accuracy, was improved by 5.2% and 3.8%, respectively, over the state-of-the-art Deeplabv3+ network. Compared to the adapted multi-task network YOLOP, MT-WavYOLO achieved a 3.4% increase in mAP50 and a 2.7% improvement in IoU. In addition, MT-WavYOLO has a compact footprint of 10.2 M parameters and achieves approximately 27 FPS in real-time inference, thereby meeting the requirements of robotic harvesting operations. The proposed localization method was evaluated through 600 fruit localization tests using six different RGB-D cameras in an orchard environment. The average experimental results demonstrated that the centroid localization and radius estimation errors were reduced by 42.5%, 73.7%, 16.17%, and 11.25%, respectively, compared to traditional 3D bounding box methods and our previous approaches. These results indicate that the MT-WavYOLO combined with the frustum-based method significantly enhances the accuracy of apple localization under complex orchard conditions using consumer-grade sensors, providing a strong practical foundation for non-destructive robotic harvesting.
Why it matches plant phenotyping methods果実の検出・3D重心定位という植物器官の形態的状態を、画像分割・深層学習・点群処理で推定する手法を開発し、データセットおよび複数カメラで性能評価しているため、方法が中心的である。
abstractwe propose a fruit detection and centroid localization method based on a Multi-Task Wavelet-Enhanced YOLO (MT-WavYOLO)
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 5 Sept 2026
The apple (Malus × domestica), a globally significant perennial fruit crop, faces immense pressure from climate change, evolving pathogens, and consumer demand for novel traits. Also, remains constrained by slow trait selection despite technological advances. Further, the traditional breeding methods are slow and resource-intensive, hampered by the apple's long juvenile period and high heterozygosity. This systematic literature review (SLR) synthesizes the state of the art in advanced data-driven technologies for accelerating apple breeding and genetic modification. Following the PRISMA-EcoEvo protocol, 47 selected studies were analyzed from databases including Web of Science, Scopus, and PubMed. Our thematic synthesis reveals a paradigm shift towards a "digital breeding" model, characterized by the convergence of three core technological pillars. First, high-throughput phenotyping (HTP), which leverages sensor modalities such as RGB-D, hyperspectral imaging, and LiDAR, is automating the collection of trait data at an unprecedented scale. Second, machine learning (ML) and deep learning (DL) algorithms are being deployed for diverse applications, including cultivar identification with over 96% accuracy, non-destructive quality prediction, and genomic selection, thereby boosting predictive ability for key traits by up to 18%. Third, precise and efficient genome editing, predominantly using Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/CRISPR-associated protein 9 (Cas9), is enabling the rapid introduction of desirable traits, such as disease resistance, enhanced shelf life, and improved nutrient uptake. Demonstrated transgene-free editing protocols are accelerating the path to commercialization. We further explore the integration of these pillars through the agricultural internet of things (AIoT) and discuss emerging frontiers, including federated learning for data privacy, explainable AI (XAI) for model transparency, and the implications of recent regulatory frameworks. This review identifies critical research gaps, including the need for standardized open-access datasets and integrated end-to-end system validation. It concludes that the synergistic application of these technologies is poised to revolutionize the speed, precision, and resilience of apple improvement programs worldwide.
Why it matches plant phenotyping methodsリンゴ育種におけるデータ駆動技術の系統的レビューであり、高スループット表現型解析のセンサー技術と技術統合・検証課題を主要に扱っている。
abstractThis systematic literature review (SLR) synthesizes the state of the art in advanced data-driven technologies for accelerating apple breeding and genetic modification.
Moisture plays a critical role in crop growth and development, making accurate, efficient, and non-destructive detection and monitoring of crop water stress essential for advancing crop science research and optimizing production management. Traditional non-destructive methods for monitoring water stress primarily rely on color imaging or partial 2D spectral analysis. However, these methods are limited to two-dimensional features and fail to capture the spatial variability of water stress within the three-dimensional canopy structure of crops. To address this limitation, this study integrates RGB-D cameras and thermal infrared cameras and introduces a method for calculating the 3D spatial distribution characteristics of crop water stress using RGB-D-T fusion analysis. This approach enables high-precision detection and analysis of water stress in strawberry plants. An RGB-D-T acquisition system was designed and implemented to collect RGB images, depth images, and thermal infrared images of strawberries subjected to different moisture gradient treatments. Using the YOLOv8-seg deep learning model, semantic segmentation of the crop canopy and the wet reference surface was performed. The segmentation results were fused with 3D point cloud data to generate a 3D dataset incorporating temperature, color, and semantic information. Subsequently, the three-dimensional distribution characteristics and dynamic changes in the canopy water stress index (CWSI) of strawberry plants were analyzed under varying moisture conditions. The results demonstrated that under low moisture gradients (15%–30%), the CWSI value increased significantly and exhibited a concentrated distribution, indicating severe water stress. Conversely, under high moisture gradients (75%–90%), the CWSI value approached zero, reflecting sufficient water supply and complete stress alleviation. Additionally, the study highlighted the variation in the temperature difference between strawberry leaves and the surrounding air, confirming the sensitivity of strawberries to water stress across different reproductive stages. The response to water deficit was most pronounced during the growth phase. By fusing multi-source data, this study achieves 3D visualization and precise quantification of water stress in strawberries, providing innovative insights and technical support for precision irrigation and crop phenotyping research.
Why it matches plant phenotyping methodsRGB-D・熱赤外センサーの融合、3D点群化、深層学習セグメンテーションにより、イチゴの水ストレスを3D定量化する取得・解析手法が研究の中心である。
abstractAn RGB-D-T acquisition system was designed and implemented to collect RGB images, depth images, and thermal infrared images of strawberries subjected to different moisture gradient treatments.
Highlights A screw-conveyor-based machine vision system was evaluated for online apple grading. The developed vision pipeline enabled multi-view-based size estimation and comprehensive surface defect inspection. Apple sizing accuracy exceeded 95.9% across conveyor speeds of 1–2 apples s-1 per conveyor lane. The system achieved sample-level grading accuracies of up to 95.4%, demonstrating its potential for further integration and in-field validation. ABSTRACT. Apple quality grading is a critical operation in postharvest handling; however, most existing grading systems are designed for controlled packinghouse environments rather than in-field operation at harvest, which can achieve substantial cost savings for both growers and packers and improve postharvest inventory management. To address the need for in-field apple grading technology, building on our prior work, this study developed a machine vision-based apple grading system toward in-field sorting by integrating a screw-conveyor-based fruit handling mechanism with automated defect inspection and size estimation. The system enables continuous fruit transportation and rotation, allowing multi-view image acquisition for comprehensive surface assessment. The vision module comprises an enclosed image chamber equipped with uniform LED illumination and a top-mounted RGB-D (red-green-blue-depth) camera, ensuring stable and consistent quality of acquired imagery. A computer vision-based pipeline was developed to detect, track, and segment individual apples for surface defect evaluation and sizing. Multi-view images acquired during fruit rotation were fused to achieve full surface coverage. In addition, a geometry-based diameter estimation method integrating stem/calyx-aware boundary localization was introduced to improve fruit sizing robustness and accuracy. Experimental results demonstrate diameter estimation accuracy exceeding 95.9% and maintained sample-level grading accuracies of 95.4%, 94.2%, and 92.3% at conveyor speeds of 1, 1.5, and 2 apples s -1 per lane, respectively. These results demonstrate that the proposed system can support rapid apple grading and provide a practical step toward in-field fruit sorting. Both the dataset and software programs of this study has been made publicly available. Keywords: Apple, In-field grading, Machine vision, Multi-view imaging, Online inspection.
Why it matches plant phenotyping methodsリンゴのサイズ推定と表面欠陥評価を行う画像ベースのオンライン表現型取得・選別システムを開発し、精度検証しているため、植物フェノタイピング手法が中心である。
abstractA computer vision-based pipeline was developed to detect, track, and segment individual apples for surface defect evaluation and sizing.
In recent years, the development of automated methods for phenotypic assessments of observable plant traits is gaining interest because they can provide advantages over standard ones. This study presents a novel methodology to estimate the canopy volume of grapevine plants during their growth cycle using a low-cost red, green, blue and depth (RGB-D) sensor mounted on a ground-based platform. The aim is to address the need for cost-effective and easy-to-use systems that can operate under variable environmental conditions without requiring highly specific data acquisition constraints and advanced technical skills. Using a Microsoft Azure Kinect RGB-D camera, detailed 3-D images of the plants are captured. The iterative closest point (ICP) algorithm is then applied to reconstruct a full view of the plants useful to estimate the canopy volumes. The effectiveness of this approach is validated by comparing the volumetric estimates with the leaf area index (LAI) measures obtained with traditional agronomic techniques during significant phenological periods of the plants’ lifecycle. The results demonstrate a correlation between the two approaches, indicating the reliability of the proposed method and highlighting advantages in terms of precision. These outcomes demonstrate the potential of automated vineyard monitoring systems, providing reliable data to assess plant growth conditions.
Why it matches plant phenotyping methodsRGB-D画像とICP再構成によりブドウ樹の樹冠体積を推定する手法を開発・検証しており、植物形質の取得が研究の中心です。
abstractThis study presents a novel methodology to estimate the canopy volume of grapevine plants during their growth cycle using a low-cost red, green, blue and depth (RGB-D) sensor mounted on a ground-based platform.
Three-dimensional (3D) reconstructions of orchards offer richer data for digital phenotyping and underpin smart-agriculture applications. However, achieving high-level reconstruction quality and robustness is challenging due to the complex structure of the orchard. This study presents a novel framework that provides centimeter-level 3D realtime reconstructions and phenotyping for apple orchards. A multi-RGBD camera array was adopted, creating a wide overlapping view and robust features. The hybrid odometry front-end and the dual loop-closure strategy ensured low-drift pose estimation. The generated pointcloud was input into the ellipsoid-fitting routine to extract fruit diameter and volume. We validated this framework through reconstruction and phenotypic errors in four rows of an apple orchard with different tree spacings. The root mean square error of the absolute trajectory error in global reconstruction was less than 16 mm. The mean absolute percentage error (MAPE) of the local fiducial distance of approximately 5 m was less than 0.12%. The system was implemented at higher than 12.5 frames per second in an embedded system. The MAPEs of the fruit’s diameter were 2–2.17%, and those of its volume were 5.3–5.6%. Additionally, ablation experiments were carried out on multi-camera and loop-closed elements, and comparisons were made with existing methods to further demonstrate their effectiveness. In conclusion, this research provides an efficient and stable deployable solution for 3D reconstruction of orchards, which is conducive to the development of more advanced and multilayer modern orchard models and promotes the practice of smart agriculture.
Why it matches plant phenotyping methodsリンゴ園向けのマルチRGB-Dによる3D再構成と、点群から果実径・体積を抽出するフェノタイピング手法を開発・検証しており、方法が研究の中心である。
abstractThis study presents a novel framework that provides centimeter-level 3D realtime reconstructions and phenotyping for apple orchards.
Controlled-environment agriculture (CEA) and circular production systems require coordinated monitoring of biological and physicochemical processes across trophic levels. This project report presents the implementation of a multi-trophic controlled-environment agriculture demonstrator that integrates computer-vision-based monitoring with established sensor infrastructure for aquaculture, poultry, plants, microalgae, duckweed, and insect modules. Stereo imaging and RGB-D systems are deployed for non-invasive quantification of fish biomass and plant growth, while continuous water-quality and environmental measurements (e.g., pH, dissolved oxygen, nitrate, ammonium, temperature, CO$_2$) provide complementary process data. These data streams are synchronized within a shared database architecture to enable cross-module evaluation of nutrient dynamics, growth progression, and operational stability under real facility conditions. The implemented framework demonstrates how computer vision can extend conventional sensor-based monitoring by directly capturing biological performance indicators across aquatic, terrestrial, and microbial domains. While advanced predictive modeling and full digital twin simulation remain future development steps, the realized data-integration architecture establishes a structural foundation for the systematic evaluation of circular indoor food-production systems. The demonstrator illustrates how multimodal monitoring can support nutrient recirculation, transparency of biological variability, and data-driven assessment within controlled multi-trophic environments.
Why it matches plant phenotyping methods植物成長をステレオ画像およびRGB-Dで非侵襲的に定量するコンピュータビジョン監視基盤を実装しており、植物フェノタイピングが統合監視システムの主要な技術要素である。
abstractStereo imaging and RGB-D systems are deployed for non-invasive quantification of fish biomass and plant growth
Potato yield is a key indicator for optimizing cultivation practices in agriculture. Potato yield can be estimated on harvesters using RGB-D cameras, which capture three-dimensional (3D) information of individual tubers moving along the conveyor belt. However, point clouds reconstructed from RGB-D images are incomplete due to self-occlusion, leading to systematic underestimation of tuber weight. To address this, we introduce PointRAFT, a high-throughput point cloud regression network that directly predicts continuous 3D shape properties, such as tuber weight, from partial point clouds. Rather than reconstructing full 3D geometry, PointRAFT infers target values directly from raw 3D data. Its key architectural novelty is an object height embedding that incorporates tuber height as an additional geometric cue, improving weight prediction under practical harvesting conditions. PointRAFT was trained and evaluated on 26,688 partial point clouds collected from 859 potato tubers across four cultivars and three growing seasons on an operational harvester in Japan. On a test set of 5,254 point clouds from 172 tubers, PointRAFT achieved a mean absolute error of 12.0 g and a root mean squared error of 17.2 g, substantially outperforming a linear regression baseline and a standard PointNet++ regression network. With an average inference time of 6.3 ms per point cloud, PointRAFT supports processing rates of up to 150 tubers per second, meeting the high-throughput requirements of commercial potato harvesters. Beyond potato weight estimation, PointRAFT provides a versatile regression network applicable to a wide range of 3D phenotyping and robotic perception tasks. The code, network weights, and a subset of the dataset are publicly available at https://github.com/pieterblok/pointraft.git.
Why it matches plant phenotyping methods部分点群からジャガイモ塊茎重量を推定する3D深層学習手法を開発・評価しており、植物形質取得が研究の中心である。
abstractwe introduce PointRAFT, a high-throughput point cloud regression network that directly predicts continuous 3D shape properties, such as tuber weight, from partial point clouds.
Reproduction assets foundThe paper publicly releases its authors' analysis code and trained network weights on GitHub, and a subset of its potato tuber partial point cloud dataset (with ground truth weights) on Hugging Face. Both are paper-specific, public, and actionable.Code · publicThe code, network weights, and a subset of the dataset are publicly available at https://github.com/pieterblok/pointraft.git .Open asset ↗pieterblok/pointraftlines:1-93Dataset · publicA subset of the datasets generated and/or analyzed during this study is publicly available at: https://huggingface.co/datasets/UTokyo-FieldPhenomics-Lab/3DPotatoTwinOpen asset ↗UTokyo-FieldPhenomics-Lab/3DPotatoTwinlines:447-463Plant phenotyping relevance match · UnverifiedOpenAlex · checked 15 Sept 2026
The operational effect of the reverse ear picking device for fresh corn is affected by stem diameter and ear orientation angle. The existing devices lack the ability to sense these parameters in real-time, making it difficult to dynamically adjust operating parameters, which leads to a high damage rate and harvest loss. To this end, this study focuses on the visual perception aspect and proposes a recognition method based on a depth camera and an improved D3-YOLOv11 segmentation model, which provides reliable visual input for subsequent adaptive regulation. Specifically, this study proposes Dual-Domain Dynamic Gate Conv (D3GConv) to enhance the multi-scale feature extraction ability of the model. In the neck network, a bidirectional weighted pyramid structure with semantic detail injection is designed to improve the segmentation accuracy of small objects. Generalized Focal Loss V2 was used to optimize the detection head to enhance the accuracy of boundary localization in dense stem scenes. Finally, the depth information is fused to realize the real-time measurement of stem diameter and ear orientation angle. Experimental results show that the Mask-mAP50 of the D3-YOLOv11 model reaches 99.3% and 94.6% in stem and ear instance segmentation tasks, respectively. The Mean Absolute Error of stem diameter measurement based on depth information is only 0.16 cm, and the Coefficient of Determination of ear orientation angle reaches 0.95, which verifies the reliability and practicability of this method in the adaptive control of the ear harvesting device. It provides an effective visual perception basis for improving the intelligence level of equipment.
Why it matches plant phenotyping methods深度カメラと改良YOLOによってトウモロコシの茎径・穂の向き角をリアルタイム測定する手法を開発・検証しており、植物形質の取得法が中心である。
abstractproposes a recognition method based on a depth camera and an improved D3-YOLOv11 segmentation model
Water availability critically affects basil (Ocimum basilicum L.) growth and physiological performance, making the early and precise monitoring of water-deficit responses essential for precision irrigation. However, conventional visual or biochemical methods are destructive and unsuitable for real-time assessment. This study presents a multimodal optical biosensing and 3D convolutional neural network (3D-CNN) fusion framework for phenotyping physiological responses of basil under water-deficit stress. RGB, depth, and chlorophyll fluorescence (CF) imaging were integrated to capture complementary morphological and photosynthetic information. Through the fusion of 130 optical parameter layers, the 3D-CNN model learned spatial and temporal–spectral features associated with resistance and recovery dynamics, achieving 96.9% classification accuracy—outperforming both 2D-CNN and traditional machine-learning classifiers. Feature-space visualization using t-SNE confirmed that the learned latent representations reflected biologically meaningful stress–recovery trajectories rather than superficial visual differences. This multimodal fusion framework provides a scalable and interpretable approach for the real-time, non-destructive monitoring of crop water stress, establishing a foundation for adaptive irrigation control and intelligent environmental management in precision agriculture.
Why it matches plant phenotyping methodsバジルの水ストレス応答を、RGB・深度・クロロフィル蛍光画像と3D-CNNで非破壊推定するフェノタイピング手法が研究の中心である。
abstractThis study presents a multimodal optical biosensing and 3D convolutional neural network (3D-CNN) fusion framework for phenotyping physiological responses of basil under water-deficit stress.
Title: ReDuCit Project Poster: Building a More Sustainable Agriculture Based on Robust, Scalable Applications of Controlled Deficit Irrigation Strategies Description: This poster presents the ReDuCit project, a 24-month initiative focused on developing sustainable irrigation strategies for citrus crops in the Guadalquivir River Basin (Spain). The project addresses the critical challenge of maintaining agricultural productivity while reducing water consumption in a region where citrus represents 5.1% of the irrigated area but accounts for 9.7% of water demand, and where climate projections indicate a 10% reduction in water availability by 2039. Project objectives include: Designing and validating an integrated system for monitoring water status and production in citrus crops Establishing a replicable and scalable model applicable to other crops Developing a digital platform for optimized irrigation management Implementing a robust Regulated Deficit Irrigation Control model capable of reducing water consumption by 15-25% Technical approach: The project combines water status monitoring (using sap flow sensors, trunk stem dendrometers, and microtensiometers on reference trees), production tracking (through autonomous RGB-D cameras with AI for fruit detection and measurement), and an integrated digital platform providing real-time data collection, automated irrigation recommendations, and personalized management alerts. Expected impact: Potential water savings of 60 million m³/year in citrus crops from the Guadalquivir region alone, representing 20% of the required reduction in the agricultural sector by 2039. Consortium: OnTech Innovation, Rovimatica, Universidad de Sevilla, Soltel Group Funding: Co-financed by European Funds through Junta de Andalucía and the Spanish Ministry of Finance Validation: Real-world testing in collaboration with the Irrigation Community of the Lower Guadalquivir Valley Bilingual poster (English/Spanish) Keywords: precision agriculture, deficit irrigation, water management, citrus crops, digital agriculture, IoT sensors, artificial intelligence, sustainability, Guadalquivir, smart farming
Why it matches plant phenotyping methods柑橘の水分状態と果実の検出・計測を行うセンサー／RGB-D・AI統合システムの設計・検証がプロジェクトの中心であり、植物状態・果実形質の取得方法を含むため。
abstractDesigning and validating an integrated system for monitoring water status and production in citrus crops
Title: ReDuCit Project Poster: Building a More Sustainable Agriculture Based on Robust, Scalable Applications of Controlled Deficit Irrigation Strategies Description: This poster presents the ReDuCit project, a 24-month initiative focused on developing sustainable irrigation strategies for citrus crops in the Guadalquivir River Basin (Spain). The project addresses the critical challenge of maintaining agricultural productivity while reducing water consumption in a region where citrus represents 5.1% of the irrigated area but accounts for 9.7% of water demand, and where climate projections indicate a 10% reduction in water availability by 2039. Project objectives include: Designing and validating an integrated system for monitoring water status and production in citrus crops Establishing a replicable and scalable model applicable to other crops Developing a digital platform for optimized irrigation management Implementing a robust Regulated Deficit Irrigation Control model capable of reducing water consumption by 15-25% Technical approach: The project combines water status monitoring (using sap flow sensors, trunk stem dendrometers, and microtensiometers on reference trees), production tracking (through autonomous RGB-D cameras with AI for fruit detection and measurement), and an integrated digital platform providing real-time data collection, automated irrigation recommendations, and personalized management alerts. Expected impact: Potential water savings of 60 million m³/year in citrus crops from the Guadalquivir region alone, representing 20% of the required reduction in the agricultural sector by 2039. Consortium: OnTech Innovation, Rovimatica, Universidad de Sevilla, Soltel Group Funding: Co-financed by European Funds through Junta de Andalucía and the Spanish Ministry of Finance Validation: Real-world testing in collaboration with the Irrigation Community of the Lower Guadalquivir Valley Bilingual poster (English/Spanish) Keywords: precision agriculture, deficit irrigation, water management, citrus crops, digital agriculture, IoT sensors, artificial intelligence, sustainability, Guadalquivir, smart farming
Why it matches plant phenotyping methods灌漑管理プロジェクトだが、果実の検出・計測を行うRGB-Dカメラ／AIと、水分状態を測定するセンサーを統合したモニタリング基盤が技術的中核として明示されており、植物の生産・生理状態の表現型取得に該当する。
abstractDesigning and validating an integrated system for monitoring water status and production in citrus crops
High-throughput phenotypic acquisition and analysis allow us to accurately quantify trait expressions, which is essential for developing intelligent breeding strategies. However, there is still much potential to explore in the field of high-throughput phenotyping for edible fungi. In this study, we developed a portable multi-phenotypic acquisition system for Pleurotus eryngii using RGB and RGB-D cameras. We developed an innovative Unet-based semantic segmentation model by integrating the ASPP structure with the VGG16 architecture. This allows for precise segmentation of the cap, gills and stem of the fruiting body. By leveraging depth images from RGB-D cameras, we can effectively collect phenotypic information about Pleurotus eryngii. By combining K-means clustering with Lab color space thresholds, we are able to achieve more precise automatic classification of Pleurotus eryngii cap colors. Moreover, AlexNet is utilized to classify the shapes of the fruiting bodies. The Aspp-VGGUnet network demonstrates remarkable performance with a mean Intersection over Union (mIoU) of 96.47% and a mean pixel accuracy (mPA) of 98.53%. These results reflect respective improvements of 3.03% and 2.23% compared to the standard Unet model, respectively. The average error in size phenotype measurement is just 0.15 ± 0.03 cm. The accuracy for cap color classification reaches 91.04%, while fruiting body shape classification achieves 97.90%. The proposed multi-phenotype acquisition system reduces the measurement time per sample from an average of 76 s (manual method) to about 2 s, substantially increasing data acquisition throughput and providing robust support for scalable phenotyping workflows in breeding research.
Why it matches plant phenotyping methodsRGB・深度画像による食用菌の複数形質取得システムを開発し、セグメンテーション、サイズ・色・形状の自動測定性能を検証しており、植物（菌類）の表現型取得が中心である。
abstractwe developed a portable multi-phenotypic acquisition system for Pleurotus eryngii using RGB and RGB-D cameras.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 6 Sept 2026
Accurate measurement of key phenotypic traits, including the horizontal and vertical diameters, the weights of both fruit and pit, is essential for the selection of elite litchi cultivars and the advancement of breeding research. Manual measurement, however, is laborious, inefficient, and subjective, highlighting the urgent need for automated and precise phenotyping tools. Unlike apples, mangoes, and grapes, litchi combines a spiny, highly variable pericarp (heterogeneous areoles/tubercles across cultivars) with diverse seed morphology (including irregular, wrinkled aborted seeds), thereby increasing the difficulty of semantic segmentation and biasing diameters and weight estimation. This study presents LitchiPhenoNet, a multimodal learning framework for litchi phenotypic analysis that employs a dual-branch architecture integrating RGB (color/texture) and depth (spatial/structural) information. Experiments were conducted on an RGB-D dataset comprising 1,198 image pairs (1280×720) across 10 cultivars, using a stratified train/test split of 958/240 pairs by cultivar. To address inherent semantic and scale inconsistencies between modalities, the framework incorporates the RD-Fusion module for precise cross-modal feature extraction, improving robustness under complex and variable pericarp surfaces. Comparative experiments show that LitchiPhenoNet consistently outperforms leading YOLO-based models, achieving millimeter-level diameter estimation with coefficients of determination approaching 0.98 and mean errors within 2 mm. For weight estimation, gram-level precision is attained across whole fruit, pit, and pulp, with coefficients of determination up to 0.98 and mean errors comparable to repeated manual measurements. By handling fine-scale surface relief and cross-cultivar variability, the framework is readily extensible to other textured fruits and scalable for high-throughput phenotyping in breeding programs. Collectively, these results demonstrate that LitchiPhenoNet provides an efficient, reliable, and accurate solution for quantifying litchi phenotypic traits, substantially advancing the objectivity and efficiency of phenotypic analysis and breeding selection.
Why it matches plant phenotyping methodsRGB-D画像を用いてライチ果実・種子・果肉の径と重量を自動推定する専用フレームワークを開発し、複数品種・比較実験で性能検証しているため、植物表現型取得法が中心である。
abstractThis study presents LitchiPhenoNet, a multimodal learning framework for litchi phenotypic analysis that employs a dual-branch architecture integrating RGB (color/texture) and depth (spatial/structural) information.
The selection of sensors for a high-throughput plant phenotyping platform is crucial for its subsequent development. It impacts the control mode, data storage and transmission, phenotype analysis algorithm and accuracy. This paper compared and evaluated the three-dimensional (3D) data acquisition performance of LiDAR, Multi-View Stereo (MVS) reconstruction, and depth image synthesis in five growth stages of maize canopies. The study found that LiDAR was the most stable and least affected by the environment. Additionally, it had the highest plant height estimation accuracy, with an average R 2 of 0.80 across all five stages. However, LiDAR is greatly affected by the stationarity of the platform and the noise of the resulting maize point cloud can be significant. The sensor required for MVS mode is low-cost, has minimal influence on platform stationarity, and allows for convenient point cloud synthesis and colour information. However, it is greatly affected by the lighting environment, resulting in a certain degree of distortion in the obtained point cloud. Additionally, it has the highest pre-processing complexity. Depth point cloud has the highest synthesis efficiency and the lowest data pre-processing complexity, making it suitable for online pre-processing and analysis. However, the initial data obtained is large and its stability is low due to its susceptibility to environmental factors. The point cloud acquired by MVS and Depth are clearer than LiDAR, making it easier for plant segmentation. This study provides a valuable foundation for the development of a high-throughput plant phenotyping platform and sensor selection.
Why it matches plant phenotyping methods植物表現型プラットフォーム向けに複数の3D取得センサーを比較評価し、トウモロコシの草丈推定精度や点群品質を検証しているため、取得・解析手法が中心です。
abstractThis paper compared and evaluated the three-dimensional (3D) data acquisition performance of LiDAR, Multi-View Stereo (MVS) reconstruction, and depth image synthesis in five growth stages of maize canopies.
Hyperspectral imaging of individual corn leaves provides valuable data for analyzing nutrient content and diagnosing diseases. However, existing leaf-level imaging techniques face challenges such as low spatial resolution and labor-intensive processes. To address these limitations, this study developed a robotic system integrated with a high-resolution line-scanning hyperspectral imaging device to autonomously scan a corn leaf. The hyperspectral imaging device used a vision-based tactile sensor for active leaf tracking throughout the scanning process, ensuring high image quality. Additionally, the device incorporated an in-hand leaf manipulation mechanism that ensured the leaf was properly positioned on the tactile sensing area at the start of every scanning. The scanning process was executed by a robotic arm equipped with an RGB-D camera and integrated with the Segment Anything Model (SAM), enabling autonomous leaf detection, localization, grasping, and scanning. The system was tested on V10-stage corn plants and the success rate was 91.4 % with an average 4.8 s for leaf detection and localization and an average leaf scanning time of 38.3 s.
Why it matches plant phenotyping methodsトウモロコシ葉の高解像度ハイパースペクトル画像取得を自動化するロボット・触覚追跡システムを開発し、検出・走査成功率や処理時間で評価しており、表現型取得手法が中心である。
abstractthis study developed a robotic system integrated with a high-resolution line-scanning hyperspectral imaging device to autonomously scan a corn leaf.
Cover crops play a critical role in providing agroecological services such as improving soil health, reducing erosion and nitrogen loss, and suppressing weeds, which are closely tied to their performance such as accumulated biomass. This study evaluated the Active Canopy Sensor (ACS) -214, an active proximal sensing device equipped with its own light-emitting red and near-infrared spectral reflectance sensors, a time-of-flight laser, and an ultrasonic sensor, for estimating winter cover crop biomass across 13 U.S. states from 2020 to 2024. We assessed 11 species from three functional groups – grasses ( n = 797), legumes ( n = 264), and brassicas ( n = 181) – using Random Forest (RF) models and four cross-validation strategies. The ACS-214 showed moderate to strong prediction accuracy for grasses ( R 2 = 0.51 – 0.64) and legumes ( R 2 = 0.44 – 0.76), though performance declined in leave-one-region-out analyses ( R 2 = 0.06 – 0.46), indicating limited spatial generalizability. Brassica models had low prediction accuracy for all models ( R 2 < 0.30), likely due to flowering and patchy growth. Biomass prediction breakpoints were observed at ∼3000 kg ha −1 for legumes and ∼4000 kg ha −1 for grasses. We also evaluated the effectiveness of using ACS-214 data to train Sentinel-2 satellite imagery for estimating grass cover crop biomass using withheld, out of bag data from 2023 to 2024. Sentinel-2 RF models trained with ACS-214 data showed good agreement with field-sampled ( R 2 = 0.58 – 0.61) and ACS-214-estimated biomass ( R 2 = 0.70). While Sentinel-2 offers scalability, the ACS-214 enables finer-resolution biomass mapping and better accounts for within-field variability, making it an effective tool for localized management and monitoring. These findings support the integration of proximal and satellite sensing approaches to enhance cover crop biomass estimation and agroecological assessment.
Why it matches plant phenotyping methods近接センサーと衛星画像を用いた作物バイオマス推定を開発・評価し、交差検証で精度と空間汎化性を検証しているため、植物形質取得法が研究の中心である。
abstractThis study evaluated the Active Canopy Sensor (ACS) -214, an active proximal sensing device equipped with its own light-emitting red and near-infrared spectral reflectance sensors, a time-of-flight laser, and an ultrasonic sensor, for estimating winter cover crop biomass across 13 U.S. states from 2020 to 2024.
Sustainable vineyard management requires precise and efficient application of plant protection products to minimise environmental impact while ensuring plant health. This study presents a variable-rate spraying system that integrates an RGB-D camera with a GPU-equipped edge computing platform to enable accurate, real-time adjustment of spray flow rates in vineyards. A tensor-based representation of RGB-D data is employed to accelerate the entire processing pipeline. Based on this structure, a fast approximate meshing method is applied to rapidly generate 3D meshes from point clouds. To incorporate semantic information from RGB images, an instance segmentation model is used to detect grapevine canopies and trellis posts. The resulting canopy masks are used to isolate the canopy meshes, while the trellis posts serve as reference planes for canopy volume estimation via mesh projection. Based on the computed volume, pulse-width modulation signals are generated to dynamically control spray flow rates. Field experiments were conducted to evaluate the system’s effectiveness and real-time performance. The results demonstrated that the estimated canopy volume is a reliable indicator for regulating application rates. Compared to uniform-rate spraying, the proposed system reduced plant protection product consumption by 57.4% while ensuring adequate droplet coverage. Additionally, the system demonstrated satisfactory real-time performance even on entry-level hardware. Overall, the proposed variable-rate spraying system offers an accurate, real-time, and cost-effective solution for precision viticulture, highlighting its potential for commercial deployment in sustainable vineyard management.
Why it matches plant phenotyping methodsRGB-D画像からブドウ樹冠を分離し、3Dメッシュ投影で樹冠体積という植物形質を推定する技術が中心であり、リアルタイム性能と散布制御への有効性も評価している。
abstractan instance segmentation model is used to detect grapevine canopies and trellis posts
Tomato is a globally significant horticultural crop with substantial economic and nutritional value. High-precision phenotypic analysis of tomato fruit characteristics, enabled by computer vision and image-based phenotyping technologies, is essential for varietal selection and automated quality evaluation. An intelligent detection framework for phenomics analysis of tomato fruits was developed in this study, which combines image processing techniques with deep learning algorithms to automate the extraction and quantitative analysis of 12 phenotypic traits, including fruit morphology, structure, color and so on. First, a dataset of tomato fruit section images was developed using a depth camera. Second, the SegFormer model was improved by incorporating the MLLA linear attention mechanism, and a lightweight SegFormer-MLLA model for tomato fruit phenotype segmentation was proposed. Accurate segmentation of tomato fruit stem scars and locular structures was achieved, with significantly reduced computational cost by the proposed model. Finally, a Hybrid Depth Regression Model was designed to optimize the estimation of optimal depth. By fusing RGB and depth information, the framework enabled efficient detection of key phenotypic traits, including fruit longitudinal diameter, transverse diameter, mesocarp thickness, and depth and width of stem scar. Experimental results demonstrated a high correlation between the phenotypic parameters detected by the proposed model and the manually measured values, effectively validating the accuracy and feasibility of the model. Hence, we developed an equipment automatically phenotyping tomato fruits and the corresponding software system, providing reliable data support for precision tomato breeding and intelligent cultivation, as well as a reference methodology for phenotyping other fruit crops.
Why it matches plant phenotyping methods深度カメラ、画像処理、深層学習を統合し、トマト果実の12形質を自動抽出・定量する装置とソフトウェアを開発しており、表現型取得法が研究の中心である。
abstractAn intelligent detection framework for phenomics analysis of tomato fruits was developed in this study, which combines image processing techniques with deep learning algorithms to automate the extraction and quantitative analysis of 12 phenotypic traits
Reproduction assets foundThe paper's Data Availability Statement explicitly deposits some datasets, model weights, and code used in the study at a public GitHub repository, which is paper-specific and actionable. Full self-developed datasets require contacting the corresponding author.Code · publicSome datasets, model weights, and code used in the present study are available at https://github.com/Snail-code-wq/Plants_Tomato_2025 (accessed on 5 November 2025). All self-developed datasets can be obtained by contacting the corresponding author.Open asset ↗Snail-code-wq/Plants_Tomato_2025lines:466-479Plant phenotyping relevance match · UnverifiedEurope PMC · OpenAlex · checked 6 Sept 2026
Abstract Purpose Dormant pruning is critical for fruit tree management and maintaining fruit quality. Traditional manual pruning is labor-intensive, driving interest in automated robotic dormant pruning. However, automated robotic dormant pruning meets significant challenges in trunk and branches detection, due to the complexity of the orchard environment and the interlacing of the branches. This paper proposes an automatic method for real-time detection and pruning of pear tree trunks and upright branches using an RGB-D camera. Methods Pear trunk detection was conducted by enhancing the You Only Look Once version 5 Nano (YOLOv5n) model with Squeeze-and-Excitation Networks (SENet) and optimizing the anchor boxes. For branch segmentation, YOLOv8n-seg integrated with Dynamic Snake Convolution (DSConv) and Focal Scale Intersection over Union (Focal_SIoU) was employed. The length and angle of the branches were calculated from the generated mask images, and the pruning position was determined. A PRUNING_ROS package was developed for real-time orchard applications. Results The evaluation demonstrated 96.7% mean average precision (mAP) for trunk detection and 82.6% mAP for branch segmentation in test dataset. Field test results showed that the mean absolute error (MAE) of trunk distance localization compared to manual measurements was 3.71 cm, with the root mean square error (RMSE) of 3.84 cm, and the frames per second (FPS) of 31.6. The MAE was 2.3 cm (RMSE: 2.6 cm) for pruning points in depth direction and 2.34° (RMSE: 2.71°) for upright branches angle, with the FPS of 36.2. The field pruning experiment showed a pruning success rate of 47.6%. Conclusion This method provides technical support for the operation of fruit tree pruning robots, representing a step toward the full automation of fruit tree management.
Why it matches plant phenotyping methodsRGB-D画像からナシ樹の幹・枝を検出・分割し、枝の長さと角度を算出する手法が中心で、単なる対象位置検出を超えた植物器官形態の計測と技術評価を行っている。
abstractFor branch segmentation, YOLOv8n-seg integrated with Dynamic Snake Convolution (DSConv) and Focal Scale Intersection over Union (Focal_SIoU) was employed. The length and angle of the branches were calculated from the generated mask images, and the pruning position was determined.
ABSTRACT RGB‐D cameras are widely used in indoor robots. However, their ranging capability for agricultural robots under natural lighting still needs to be evaluated. Especially in the field of robotics‐based high‐throughput crop phenotyping, the measurement accuracy of phenotypic parameters is deeply related to the ranging performances of RGB‐D cameras. In this paper, we propose a depth‐ranging evaluation framework and an online ranging compensation strategy for RGB‐D cameras on phenotyping robots. The goal is to acquire high‐quality depth‐ranging performances for plant phenotyping tasks. First, we evaluate ranging performances of RealSense D435i and Kinect V2 under typical phenotyping scenes with different lighting conditions, verify their feasibility on different maize organ observations in different growth periods, and give the optimal observation ranging areas. Second, we employ image brightness to reflect the lighting situations, and propose a novel ranging compensation strategy to decrease the lighting influences in real‐time. The results of sufficient field experiments show that RealSense D435i has better ranging performances than Kinect V2 for crop phenotyping, especially for open‐field, in‐row, and close‐range observations. The optimal ranging area of RealSense D435i is within a region of [0.16–1.2] m. However, Kinect V2 is not suitable for field phenotyping robots due to significant interference from natural sunlight, limited measurement range, and instability in depth measurements under outdoor conditions. In addition, we also verify that our online depth error compensation strategy can effectively reduce the influences of lighting intensity and target distance on the depth ranging of RGB‐D cameras. Although we test and verify our ranging evaluation framework and ranging error compensation strategy with two old‐fashion cameras, the framework and strategy are generic and applicable to other new RGB‐D cameras.
Why it matches plant phenotyping methods植物フェノタイピングロボット向けRGB-Dカメラの測距評価フレームワークとオンライン補償手法を開発・検証しており、表現型取得の技術性能が研究の中心である。
abstractwe propose a depth‐ranging evaluation framework and an online ranging compensation strategy for RGB‐D cameras on phenotyping robots.
An autonomous IoT crop phenotyping system has been developed that integrates microclimatic, optical, and soil measurements with low-energy LoRaWAN connectivity. The ESP32-S3 node, lifepower, and periodic surveys (1 hour) ensure long-term operation. The optical module is based on the AS7262/AS7263 Fresnel lens spectrometers; the PHAR metrics are validated relative to the LI-190SB quantum sensor. According to field measurements in 2024. The diurnal profiles match, and the spectral features - PPFD regression model explains 89% of the variance (R2=0.89). The three-block architecture (aboveground/underground/control) is complemented by a modular infrared CO2 gas analyzer and a ToF laser sensor for calculating plant biomass growth and potential prediction of phenophases. It is shown that an inexpensive sensor assembly provides a reproducible assessment of biophysical parameters sufficient for rapid diagnosis of crop heterogeneity and subsequent integration with productivity models. Keywords: PHENOTYPING, INTERNET OF THINGS, IoT, AGROECOLOGICAL MONITORING, PRECISION AGRICULTURE, CROP HETEROGENEITY, REMOTE SENSING, LORAWAN, PAR
Why it matches plant phenotyping methods植物の生育・バイオマス・フェノフェーズ等を取得するIoTセンシング基盤の開発と、量子センサーとの検証が中心である。
abstractAn autonomous IoT crop phenotyping system has been developed that integrates microclimatic, optical, and soil measurements with low-energy LoRaWAN connectivity.
Amid growing challenges to global food security, high-throughput crop phenotyping has become an essential tool, playing a critical role in genetic improvement, biomass estimation, and disease prevention. Unlike controlled laboratory environments, field-based phenotypic data collection is highly vulnerable to unpredictable factors, significantly complicating the data acquisition process. As a result, the choice of appropriate data collection equipment and processing methods has become a central focus of research. Currently, three key technologies for extracting crop phenotypic parameters are Light Detection and Ranging (LiDAR), Multi-View Stereo (MVS), and depth camera systems. LiDAR is valued for its rapid data acquisition and high-quality point cloud output, despite its substantial cost. MVS offers the potential to combine low-cost deployment with high-resolution point cloud generation, though challenges remain in the complexity and efficiency of point cloud processing. Depth cameras strike a favorable balance between processing speed, accuracy, and cost-effectiveness, yet their performance can be influenced by ambient conditions such as lighting. Data processing techniques primarily involve point cloud denoising, registration, segmentation, and reconstruction. This review summarizes advances over the past five years in 3D reconstruction technologies—focusing on both hardware and point cloud processing methods—with the aim of supporting efficient and accurate 3D phenotype acquisition in high-throughput crop research.
Why it matches plant phenotyping methods作物キャノピーの3D形質取得に用いるLiDAR、MVS、深度カメラと点群処理を中心に扱うレビューであり、植物フェノタイピング手法が主題。
titleApplications of 3D Reconstruction Techniques in Crop Canopy Phenotyping: A Review
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.
Reproduction assets foundThe authors publicly release the paper's lychee RGB-D image dataset (raw/augmented RGB images, depth maps, detection and maturity annotations) and the Python scripts for data augmentation, image similarity comparison, and annotation in the same GitHub repository.Dataset · publicchees, the non-augmented models
produced misclassifications with lower recognition and accuracy, whereas the augmented models
avoided these issues. Overall, the results demonstrate that the data augmentation method effectively
improves the comprehensive performance of the models.
5. Data Availability
The dataset is available at：https://github.com/SeiriosLab/Lychee. The Python scripts for data
augmentation, image similarity comparison, and annotation are available within the same
repository under the tree/main/script directory.Open asset ↗SeiriosLab/Lycheepdf-raw-page:13 lines:1-55Code / dataset availability confirmedOpenAlex · Europe PMC · checked 14 Sept 2026
Accurate 3D phenotyping of agricultural produce remains challenging due to the trade-off between reconstruction quality and acquisition throughput in existing sensing technologies. While RGB-D cameras enable high-throughput scanning in operational settings like harvesting conveyors, they produce incomplete, low-quality 3D models. Conversely, close-range Structure-from-Motion (SfM) produces high-quality reconstructions but is not suitable for high-throughput field application. This study bridges this gap through 3DPotatoTwin , a paired dataset containing 339 tuber samples across three cultivars collected in Hokkaido, Japan. Our dataset uniquely combines: (1) conveyor-acquired RGB-D point clouds, (2) ground measurement, (3) SfM reconstructions under indoor controlled environment, and (4) aligned model pairs with transformation matrices. The multi-sensory alignment employs an semi-supervised pin-guided pipeline incorporating single-pin extraction and referencing, cross-strip matching, and binary-color-enhanced ICP, achieving 0.59 ± 0.11 mm registration accuracy. Beyond serving as a benchmark for 3D phenotyping algorithms, the dataset enables training of 3D completion networks to reconstruct high-quality 3D models from partial RGB-D point clouds. Meanwhile, the proposed semi-automated annotation pipeline has the potential to accelerate 3D dataset generation for similar studies. The presented methodology demonstrates broader applicability for multi-sensor data fusion across crop phenotyping applications. The dataset and pipeline source code are publicly available at HuggingFace and GitHub, respectively.
Why it matches plant phenotyping methodsジャガイモ塊茎の3D表現型計測を対象に、RGB-D・SfM・地上計測を統合したデータセット、位置合わせパイプライン、ベンチマークを開発しており、表現型取得手法が中心である。
abstractAccurate 3D phenotyping of agricultural produce remains challenging due to the trade-off between reconstruction quality and acquisition throughput in existing sensing technologies.
Reproduction assets found保存済みの本文根拠を更新済みルールで再検証し、公開資産1件を確認しました。Code · publicAll the batch processing scripts mentioned in this section were provided in the 3dscan folder at Github (https://github.com/UTokyo-FieldPhenomics-Lab/PotatoScan/).Open asset ↗UTokyo-FieldPhenomics-Lab/PotatoScanhtml-lines:119-131Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Published1 Oct 2025Computers and Electronics in Agriculture.
TomatoGreenhouseLiDAR / point cloudRGB-D / ToFStem / branch2D/3D reconstructionArchitecture / morphology / geometryGrowth / development / phenology
Growth monitoring of tomato plants in large greenhouse environments is critical for quality and efficient production. Stem diameter and elongation are key phenotypic traits for plant growth monitoring. Traditional methods, however, rely on manual operations, which are time-consuming and labor-intensive and do not apply to large-scale greenhouses. Currently, automated image-based methods exemplified by three-dimensional (3D) point cloud technology are among the preferred solutions. Nevertheless, the occlusion of plant structures during the information acquisition process is challenging for practical applications. To address this challenge, this study proposes a novel method for plant stem occlusion inpainting using Deep Reinforcement Learning (DRL). Unlike most existing 3D reconstruction approaches that require depth data from multiple viewpoints, our solution captures 3D point cloud data from a single direction. The DRL model is applied to inpaint the incomplete stem for accurate stem reconstruction and phenotypic measurements. Specifically, our approach consists of two parts, structural completion and stem diameter completion. First, we extract the point cloud of incomplete stems from the RGB-D camera data. Second, we obtain the spatial structure of the stems by inpainting the 3D stem centerline with the DRL model. Finally, we add shape features (stem diameters) by inpainting the two edge lines of the stem occlusion part with the DRL model. For stem inpainted 3D point cloud data, we conducted validation experiments by measuring several commonly used stem phenotypic traits in tomato plants, including stem diameter, stem length, and stem inclination. The experimental results show that the Mean Absolute Percentage Error (MAPE) of the occluded main stem diameter is 9.7%, stem length is 5.7%, and tilt angle is 1%. For the occluded branch stem, the MAPE of stem diameter is 23.1%, stem length is 7.9%, and tilt angle is 1.5%. The accuracy of these measurements for occluded stems is acceptable compared to that obtained from 3D point clouds of unoccluded stems. This highlights the significant potential of using DRL to effectively inpaint occluded 3D point cloud data of plants.
Why it matches plant phenotyping methods植物茎の遮蔽部分を3D点群と深層強化学習で補完し、茎径・茎長・傾斜角という表現型形質の測定精度を検証しており、フェノタイピング手法が中心的です。
abstractTo address this challenge, this study proposes a novel method for plant stem occlusion inpainting using Deep Reinforcement Learning (DRL).
Accurate estimation of comprehensive traits such as yield and quality is crucial for optimizing agricultural management practices across the tomato industry chain. Traditional manual methods are time-consuming, labor-intensive, and prone to errors, reducing estimation accuracy. In contrast, modern intelligent estimation approaches based on multi-temporal spatial and spectral feature fusion offer improved efficiency and accuracy but still face challenges such as non-generalizable segmentation models, asynchronous feature extraction and weak correlations. This study proposes a novel pipeline for estimating yield and quality of greenhouse tomatoes using temporal semantic multispectral (TSM) point clouds. An unsupervised deep learning model was designed to register RGB-D images and multispectral (MS) images collected by an unmanned ground vehicle (UGV) plant phenotyping platform. The digital number (DN) point clouds of tomato organs were reconstructed based on the masks predicted by SegFormer with fusion of multispectral and depth modalities (MSD-SF). These point clouds were then radiometrically calibrated using neural reference field with sparse viewpoints (NeREF-S) to generate accurate reflectance point clouds. Finally, multi-temporal spatial-spectral features of tomatoes were extracted from the TSM point clouds, and random forest regression models were developed to estimate traits such as fruit flavor preference, water content, brix, acidity, brix-to-acid ratio, vitamin C content, single-fruit mass, and single-plant yield. The image registration model achieved high accuracy on the test set, with average structural similarity index measure, peak signal-to-noise ratio and learned perceptual image patch similarity of 0.238, 13.116 dB, and 0.374, respectively. The MS point clouds calibrated by NeREF-S significantly improved the signal-to-noise ratio to 11.56 dB. The average rRMSE for all trait estimations was 9.03 %. The results indicate that the proposed estimation method is efficient and accurate, holding promise to become a new paradigm for estimating the comprehensive traits of greenhouse tomatoes.
Why it matches plant phenotyping methods温室トマトの収量・品質形質を推定するため、UGVフェノタイピングプラットフォーム、マルチスペクトル点群生成、画像登録・放射較正、特徴抽出および回帰推定パイプラインを中心的に開発・評価している。
abstractThis study proposes a novel pipeline for estimating yield and quality of greenhouse tomatoes using temporal semantic multispectral (TSM) point clouds.
The use of multiple camera technologies in a combined multimodal monitoring system for plant phenotyping offers promising benefits. Compared to configurations that rely on a single camera technology, cross-modal patterns can be recorded that allow a more comprehensive assessment of plant phenotypes. However, the effective utilization of cross-modal patterns depends on image registration to achieve pixel-precise alignment - a challenge often complicated by parallax and occlusion effects inherent in plant canopy imaging. In this study, we propose a novel multimodal 3D image registration method that addresses these challenges by integrating depth information from a time-of-flight camera into the registration process. By leveraging depth data, our method mitigates parallax effects, facilitating more accurate pixel alignment across camera modalities. Additionally, we introduce an automated mechanism to identify and differentiate various types of occlusions, thereby minimizing registration errors. To evaluate the efficacy of our approach, we conduct experiments on a diverse dataset comprising six distinct plant species with varying leaf geometries. Our results demonstrate the robustness of the proposed registration algorithm, showcasing its ability to achieve accurate alignment across different plant types and camera compositions. Compared to previous methods our approach is not reliant on detecting plant-specific image features, making it suitable for a wide range of applications in plant sciences. Moreover, the registration approach can scale to arbitrary numbers of cameras with varying resolutions and wavelengths. Overall, our study contributes to advancing the field of plant phenotyping by offering a robust and reliable solution for multimodal image registration.
Why it matches plant phenotyping methods植物フェノタイピング向けのマルチモーダル3D画像位置合わせ手法を開発し、複数植物種の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。
abstractIn this study, we propose a novel multimodal 3D image registration method that addresses these challenges by integrating depth information from a time-of-flight camera into the registration process.
Aiming to address the accuracy problem of cane tip recognition in complex natural environments, this paper proposes a cane tip feature annotation method based on the growth characteristics of sugarcane. In the context of the demand for lightweight and fast detection of cane tips, this paper optimizes the Yolov8n-Seg model with lightweight shared convolutional separated batch normalized detection head, model pruning, and knowledge distillation strategies. With these improvements, the accuracy of the optimized model increased by 0.2 percentage points, the number of parameters was reduced by 75.03%, the model size was reduced by 70.15%, the inference time is accelerated by 17.34%, and the GFLOPs were reduced by 40.00%. The lightweight cane tip detection model was deployed on the Jetson Orin NX platform with an average recognition frame rate of 7.42 f/s provides a lightweight hardware deployment solution for real-world applications in sugarcane harvesters. Finally, the depth camera was used for cane tip recognition and height measurement. The experimental results showed that the average relative errors of the camera were 0.189%, 0.675%, and 0.949% when the camera was 50 cm, 75 cm, and 100 cm away from the cane tip, respectively, which were all controlled within 1%, and were able to achieve accurate height measurement. Based on the statistical analysis of sugarcane clusters, this paper further proposes a sugarcane cluster identification method, providing a theoretical basis for saving adjustment time of the tip cutter during the harvesting process. It lays a theoretical and technical foundation for researching feature recognition, cutter height positioning, and real-time control of sugarcane harvester cuttings.
Why it matches plant phenotyping methodsサトウキビ先端の画像認識と深度カメラによる高さ測定を開発・検証しており、植物形態形質の取得が中心である。
abstractthis paper proposes a cane tip feature annotation method based on the growth characteristics of sugarcane.
Grapevine winter pruning is a labor-intensive and repetitive process that significantly influences grape yield and quality at harvest and produced wine. Due to its complexity and repetitive nature, the task demands skilled labor that needs to be trained, as in many other agricultural sectors. This paper encompasses an approach that targets using a robotic system to perform autonomous grapevine winter pruning using a vision system and artificial intelligence. In our previous work, we presented a 2D neural network that segmented images of grapevines into 5 different classes of plant organs during their dormant season. In this paper, we expand into the third dimension, introducing point clouds into our algorithm. The 3D approach creates instance-segmented point clouds using depth images and segmentation masks obtained with our 2D neural network. After the 3D reconstruction, the system extracts thickness measurement and uses agronomic knowledge to place pruning points for balanced pruning. The study not only delineates the integration of 2D and 3D methods but also scrutinizes their efficacy in pruning point identification. The real-world performance of the created system was evaluated and statistically analyzed on data collected during field trials in the winter pruning season 2022/2023, where the system was used in a potted vineyard to prune a set of test vines, where the positive success rate is 54.2%. Moreover, as one of the main contributions, the paper underscores a unique facet of adaptability, presenting a customizable framework that empowers end-users to fine-tune parameters according to the expected balanced pruning. This adaptability extends to variables such as the number of nodes to retain on pruned spurs and the preferred cane thickness, encapsulating the versatility of the 3D approach.
Why it matches plant phenotyping methods2D画像分割と3D点群再構成を統合し、ブドウ樹器官の厚さを抽出して剪定点を生成・評価する手法が研究の中心であり、植物形質の取得と技術性能検証を含む。
abstractThe 3D approach creates instance-segmented point clouds using depth images and segmentation masks obtained with our 2D neural network.
The use of multiple camera technologies in a combined multimodal monitoring system for plant phenotyping offers promising benefits. Compared to configurations that rely on a single camera technology, cross-modal patterns can be recorded that allow a more comprehensive assessment of plant phenotypes. However, the effective utilization of cross-modal patterns depends on image registration to achieve pixel-precise alignment - a challenge often complicated by parallax and occlusion effects inherent in plant canopy imaging. In this study, we propose a novel multimodal 3D image registration method that addresses these challenges by integrating depth information from a time-of-flight camera into the registration process. By leveraging depth data, our method mitigates parallax effects, facilitating more accurate pixel alignment across camera modalities. Additionally, we introduce an automated mechanism to identify and differentiate various types of occlusions, thereby minimizing registration errors. To evaluate the efficacy of our approach, we conduct experiments on a diverse dataset comprising six distinct plant species with varying leaf geometries. Our results demonstrate the robustness of the proposed registration algorithm, showcasing its ability to achieve accurate alignment across different plant types and camera compositions. Compared to previous methods our approach is not reliant on detecting plant-specific image features, making it suitable for a wide range of applications in plant sciences. Moreover, the registration approach can scale to arbitrary numbers of cameras with varying resolutions and wavelengths. Overall, our study contributes to advancing the field of plant phenotyping by offering a robust and reliable solution for multimodal image registration.
Why it matches plant phenotyping methods植物フェノタイピング向けのマルチモーダル3D画像位置合わせ手法を開発・評価しており、表現型取得ワークフローの技術的中心である。
abstractwe propose a novel multimodal 3D image registration method that addresses these challenges by integrating depth information from a time-of-flight camera into the registration process.
• RGB-D camera and Jetson platform enable precise, adaptive vineyard spraying. • Canopy volume estimation reduces plant protection product use by 57.4%. • Real-time system adjusts spray rates for efficient, sustainable vineyard management. • Instance segmentation and 3D meshing provide accurate canopy and trellis detection. • Jetson’s parallel computing accelerates processing for fast, reliable results. Sustainable vineyard management requires precise and efficient application of plant protection products to minimise environmental impact while ensuring plant health. This study presents a variable-rate spraying system that integrates an RGB-D camera with a GPU-equipped edge computing platform to enable accurate, real-time adjustment of spray flow rates in vineyards. A tensor-based representation of RGB-D data is employed to accelerate the entire processing pipeline. Based on this structure, a fast approximate meshing method is applied to rapidly generate 3D meshes from point clouds. To incorporate semantic information from RGB images, an instance segmentation model is used to detect grapevine canopies and trellis posts. The resulting canopy masks are used to isolate the canopy meshes, while the trellis posts serve as reference planes for canopy volume estimation via mesh projection. Based on the computed volume, pulse-width modulation signals are generated to dynamically control spray flow rates. Field experiments were conducted to evaluate the system’s effectiveness and real-time performance. The results demonstrated that the estimated canopy volume is a reliable indicator for regulating application rates. Compared to uniform-rate spraying, the proposed system reduced plant protection product consumption by 57.4% while ensuring adequate droplet coverage. Additionally, the system demonstrated satisfactory real-time performance even on entry-level hardware. Overall, the proposed variable-rate spraying system offers an accurate, real-time, and cost-effective solution for precision viticulture, highlighting its potential for commercial deployment in sustainable vineyard management.
Why it matches plant phenotyping methodsRGB-D画像、インスタンスセグメンテーション、3Dメッシュによりブドウ樹冠を抽出し、樹冠体積という植物形質を推定する手法とリアルタイム基盤が研究の中心であるため。
abstractThis study presents a variable-rate spraying system that integrates an RGB-D camera with a GPU-equipped edge computing platform to enable accurate, real-time adjustment of spray flow rates in vineyards.
Obtaining the complete morphology of tomato fruits under non-destructive conditions is essential for phenotype research, yet fruit occlusions often hinder deep learning-based image segmentation methods from capturing the true shape of occluded regions. This limitation reduces prediction accuracy and adversely impacts phenotype data acquisition. To overcome this challenge, we propose CGA-ASNet, an RGB-D amodal segmentation network incorporating a Contextual and Global Attention (CGA) module. A synthetic tomato dataset (Tomato-sim) was constructed using NVIDIA Isaac Sim's Replicator Composer (ISRC) to realistically simulate tomato morphology and greenhouse environments, and the network was trained on this dataset. To evaluate generalization, CGA-ASNet was tested on both the synthetic and a separate real-world dataset. While no explicit domain adaptation techniques were adopted, diverse lighting conditions (strong, normal, and weak illumination) were simulated to implicitly reduce the domain gap, and a mean coordinate fusion algorithm was introduced to improve annotation completeness in real-world occlusion scenarios. By leveraging contextual information among feature input keys for self-attention learning, capturing global information, and expanding the receptive field, CGA-ASNet enhanced representation capacity, semantic understanding, and localization accuracy. Experimental results demonstrated that CGA-ASNet achieved an F@0.75 score of 94.2 and a mean Intersection over Union (mIoU) of 82.4% in greenhouse amodal segmentation tasks. These findings indicate that training with well-designed synthetic datasets can effectively support accurate occlusion-aware segmentation in real environments, providing a practical solution for tomato phenotyping in greenhouse conditions.
Why it matches plant phenotyping methodsトマト果実の遮蔽領域を復元して完全形態を取得するRGB-D画像解析手法を開発し、合成・実画像データセットで技術検証しているため、植物表現型取得が中心的である。
abstractObtaining the complete morphology of tomato fruits under non-destructive conditions is essential for phenotype research
Pomegranate (Punica granatum) fruit size estimation plays a crucial role in orchard management decision-making, especially for fruit quality assessment and yield prediction. Currently, fruit sizing for pomegranates is performed manually using calipers to measure equatorial and polar diameters. These methods rely on human judgment for sample selection, they are labor-intensive, and prone to errors. In this work, a novel framework for automated on-tree detection and sizing of pomegranate fruits by a farmer robot equipped with a consumer-grade RGB-D sensing device is presented. The proposed system features a multi-stage transfer learning approach to segment fruits in RGB images. Segmentation results from each image are projected on the co-located depth image; then, a fruit clustering and modeling algorithm using visual and depth information is implemented for fruit size estimation. Field tests carried out in a commercial orchard are presented for 96 pomegranate fruit samples, showing that the proposed approach allows for accurate fruit size estimation with an average discrepancy with respect to caliper measures of about 1.0 cm on both the polar and equatorial diameter.
Why it matches plant phenotyping methodsRGB-D画像と深度情報を用いて樹上果実を検出・セグメント化し、果実サイズを推定する手法が研究の中心であり、キャリパー測定による技術検証も行っているため。
abstracta novel framework for automated on-tree detection and sizing of pomegranate fruits by a farmer robot equipped with a consumer-grade RGB-D sensing device is presented
Accurate segmentation of key tobacco structures is essential for enabling automated harvesting. However, complex backgrounds, variable lighting conditions, and blurred boundaries between the stem and petiole significantly hinder segmentation accuracy in field environments. To overcome these challenges, we propose an enhanced instance segmentation approach based on YOLOv8-seg, incorporating depth-based background filtering and architectural improvements. Specifically, depth information from RGB-D images is employed to spatially filter non-target background regions, thereby enhancing foreground clarity. In addition, a Hybrid Dilated Residual Attention Block (HDRAB) is integrated into the YOLOv8-seg backbone to improve boundary discrimination between petioles and stems, while a Lightweight Shared Detail-Enhanced Convolution Detection Head (LSDECD) is designed to efficiently capture fine-grained texture features. Experimental results demonstrate that depth filtering increases mAP50 bb and mAP50 seg by 7.9% and 6.3%, respectively, while the architectural enhancements further raise them to 89.5% and 91.1%, surpassing the YOLOv8-seg baseline by 5.2% and 10.0%. Compared with mainstream models such as Mask R-CNN and SOLOv2, the proposed method achieves superior segmentation accuracy with low computational cost, highlighting its potential for practical deployment in automated tobacco harvesting.
Why it matches plant phenotyping methodsタバコ植物の茎・葉柄などの構造をRGB-D画像からセグメンテーションする手法を開発・比較しており、植物器官形態の取得が中心である。
abstractwe propose an enhanced instance segmentation approach based on YOLOv8-seg, incorporating depth-based background filtering and architectural improvements.
Accurate, high-throughput phenotyping is a critical component of modern crop breeding programs, especially for improving traits such as mechanical stability, biomass production, and disease resistance. Stalk diameter is a key structural trait, but traditional measurement methods are labor-intensive, error-prone, and unsuitable for scalable phenotyping. In this paper, we present a geometry-aware computer vision pipeline for estimating stalk diameter from RGB-D imagery. Our method integrates deep learning-based instance segmentation, 3D point cloud reconstruction, and axis-aligned slicing via Principal Component Analysis (PCA) to perform robust diameter estimation. By mitigating the effects of curvature, occlusion, and image noise, this approach offers a scalable and reliable solution to support high-throughput phenotyping in breeding and agronomic research.
Why it matches plant phenotyping methodsRGB-D画像から作物の茎径を推定するコンピュータビジョン手法を開発しており、植物形質の取得・抽出が研究の中心であるため。
abstractIn this paper, we present a geometry-aware computer vision pipeline for estimating stalk diameter from RGB-D imagery.
In situ detection of growth information in greenhouse crops is crucial for germplasm resource optimization and intelligent greenhouse management. To address the limitations of poor flexibility and low automation in traditional phenotyping platforms, this study developed a controlled environment inspection robot. By means of a SCARA robotic arm equipped with an information acquisition device consisting of an RGB camera, a depth camera, and an infrared thermal imager, high-throughput and in situ acquisition of lettuce phenotypic information can be achieved. Through semantic segmentation and point cloud reconstruction, 12 phenotypic parameters, such as lettuce plant height and crown width, were extracted from the acquired images as inputs for three machine learning models to predict fresh weight. By analyzing the training results, a Backpropagation Neural Network (BPNN) with an added feature dimension-increasing module (DE-BP) was proposed, achieving improved prediction accuracy. The R2 values for plant height, crown width, and fresh weight predictions were 0.85, 0.93, and 0.84, respectively, with RMSE values of 7 mm, 6 mm, and 8 g, respectively. This study achieved in situ, high-throughput acquisition of lettuce phenotypic information under controlled environmental conditions, providing a lightweight solution for crop phenotypic information analysis algorithms tailored for inspection tasks.
Why it matches plant phenotyping methods温室内ロボット、複数センサー、画像解析、形質抽出、重量推定を一体化した植物表現型取得手法の開発が中心である。
abstractthis study developed a controlled environment inspection robot
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.
Reproduction assets foundThe paper's data availability statement explicitly links a public GitHub repository containing part of the data and code supporting this PhenoRob-F phenotyping study; remaining data are available on request.Code · publicPart of the data and code supporting this study are openly available with the following link: https://github.com/balloonhaha/PhenoRob-F. All other reasonable requests for data and research materials will be fulfilled upon contacting the corresponding authors.Open asset ↗balloonhaha/PhenoRob-Fhtml-lines:193-220Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Published1 Sept 2025Computers and Electronics in Agriculture.
Harvesting is a critical task in the tree fruit industry, demanding extensive manual labor and substantial costs, and exposing workers to potential hazards. Recent advances in automated harvesting offer a promising solution by enabling efficient, cost-effective, and ergonomic fruit picking within tight harvesting windows. However, existing harvesting technologies often indiscriminately harvest all visible and accessible fruits, including those that are unripe or undersized. This study introduces a novel foundation-model-based framework for efficient apple ripeness and size estimation. Specifically, we curated two public RGBD-based Fuji apple image datasets, integrating expanded annotations for ripeness (“Ripe” vs. “Unripe”) based on fruit color and image capture dates. The resulting comprehensive dataset, Fuji-Ripeness-Size Dataset, includes 4,027 images and 16,257 annotated apples with ripeness and size labels. To the best of our knowledge, this is the first published dataset on apples with ripeness and size annotations. Leveraging Grounding-DINO, a foundation-model-based object detector, we achieved robust apple detection and ripeness estimation, with mean Average Precision being 72.8, outperforming other state-of-the-art models in the evaluation on our dataset. Additionally, we developed six size estimation algorithms, made a comprehensive comparison using box-plots, and identified the best algorithm with lowest error and variation. The Fuji-Ripeness-Size Dataset and the apple detection and size estimation algorithms are made publicly available¹1The code and dataset is available at https://github.com/zhukeyi-stan/Fuji_Ripeness_And_Size_Estimation., which provides valuable benchmarks for future studies in automated and selective harvesting.
Why it matches plant phenotyping methodsリンゴの熟度・サイズという植物器官の形質を画像から推定する手法を開発・比較し、データセットとベンチマークも提供しているため、フェノタイピング手法が中心である。
abstractThis study introduces a novel foundation-model-based framework for efficient apple ripeness and size estimation.
Thinning is a critical practice in apple orchard management, directly influencing crop load and fruit quality. To assist automated crop load management, a machine vision system for apple bud detection was developed to be integrated with robotic platforms. The system employed a Kinect Azure sensor for real-time bud detection and branch diameter measurement, utilizing a YOLOv8-based object detection model trained and evaluated across multiple datasets. The evaluation identified the best-performing model by balancing precision, recall, and robustness in the complex and unstructured environments of apple orchards. Several training configurations were assessed, with the selected setup demonstrating a strong balance between precision (68 %), recall (55 %), F1-score (61 %), and mean average precision (mAP: 59 %) across diverse and unstructured orchard environments. This configuration, trained on a combination of FLIR and Kinect Azure data, was chosen for deployment due to its robustness and compatibility with the Kinect Azure sensor in real-world applications. Two proposed imaging methods for branch diameter measurement were validated against manual caliper-based measurements, with statistical analysis revealing no significant differences (p = 0.98). These findings confirm the semi-automated methods as reliable and labor-efficient alternatives for field applications. Additionally, the bud counting algorithm demonstrated accurate tracking and counting of apple buds, effectively avoiding omissions and duplications in real orchard settings. This study underscores the potential of vision systems to revolutionize apple bud thinning, providing a strong foundation for the development of fully automated solutions in precision orchard management.
Why it matches plant phenotyping methodsリンゴ芽の画像検出に加え、枝径という植物形質の画像計測法を開発・手動測定と検証しており、フェノタイピング手法が中心である。
abstracta machine vision system for apple bud detection was developed to be integrated with robotic platforms
This study proposes a novel system for estimating the 3D structure of grapevines as part of a robotic pruning system. The system aims to accurately and efficiently estimate grapevine structures. Utilizing two RGB-D cameras based on Time-of-Flight (ToF) technology, depth images were captured from a wide field of view. This paper employs a minimum spanning tree (MST) to estimate the grapevine skeleton using a cost function that considers node distance, gravitropism, and connection smoothness. Notably, we developed a new component classification method that accurately classifies structural parts—cordons, shoots, and buds—using only skeletal information. The results demonstrated that the system could effectively distinguish between different parts of the grapevine using just the 3D skeletal structure. The system was evaluated on 10 grapevines in real-world vineyard environments. The proposed method achieved high alignment accuracy with manually constructed ground-truth skeletons, even under occlusion. For bud estimation, statistical analysis based on field data facilitated effective estimation. The proposed method outperformed existing approaches in processing speed, with an average processing time of 619 ms per grapevine. These results indicate the potential for real-time application in robotic pruning, enabling efficient structure estimation with high accuracy. Future work will focus on integration with other subsystems integrated within a robotic pruning system, which is expected to produce synergistic effects and enhance overall system performance.
Why it matches plant phenotyping methodsRGB-D画像からブドウ樹の3D構造、器官分類、芽数を推定する手法を開発・評価しており、ロボット剪定への応用を超えて植物形態フェノタイピングが中心である。
abstractThis study proposes a novel system for estimating the 3D structure of grapevines as part of a robotic pruning system.
Yield mapping in agricultural crops remains a significant challenge, particularly in uncontrolled environments. This study evaluates four instance segmentation algorithms: YOLOv8n, YOLOv8s, YOLOv8m, and YOLOv8l, along with a low-cost GNSS RTK system to detect and count strawberries in a hydroponic environment. A depth camera is used to remove background information from nonrelevant furrows, improving fruit detection accuracy. The low-cost RTK receiver, configured in Base Rover mode, provides centimeter-level precision and enables the generation of detailed yield maps that can be seamlessly integrated into commercial systems to increase growers' yield and profit. Data were collected in the municipality of Arcabuco, Boyacá (Colombia), resulting in 8848 images processed after augmentation. Among the models evaluated, YOLOv8l achieved the highest performance with a maximum F1-Score of 0.9295 and a mAP50 of 0.9689 during validation. Furthermore, in the fruit counting process - evaluated against manual counts - the same model achieved a R 2 of 0.9997 and a mean relative error (MRE) of 1.5511%. In general, this work presents a systematic methodology for the extraction and visualization of information in fruit crops using computer vision and deep learning, showcasing a robust yield mapping system. The approach integrates pre-processing and post-processing steps, as well as 2D–3D image acquisition, georeferencing, and processing technologies, offering thus a novel solution for accurate and efficient hydroponic strawberry yield mapping. • Evaluation of instance segmentation to detect and count hydroponic strawberries. • 3D camera integration to remove background noise and improve fruit detection. • Use of GNSS RTK with centimeter accuracy for fruit localization in yield maps. • Real-world yield mapping system showing exact strawberry count and location. • Pre/postprocessing and georeferencing for a yield mapping approach in strawberries.
Why it matches plant phenotyping methodsイチゴ果実の検出・計数という植物器官の収量形質を、RGB-D画像、深層学習、GNSS RTKで取得・抽出・地図化する方法が研究の中心である。
abstractThis study evaluates four instance segmentation algorithms: YOLOv8n, YOLOv8s, YOLOv8m, and YOLOv8l, along with a low-cost GNSS RTK system to detect and count strawberries in a hydroponic environment.
For fruit harvesting robots, rapid and accurate detection of fruits and picking points is one of the main challenges for their practical deployment. Several fruits typically grow in clusters or bunches, such as grapes, cherry tomatoes, and blueberries. For such clustered fruits, it is desired for them to be picked by bunches instead of individually. This study proposes utilizing a low-cost off-the-shelf RGB-D sensor mounted on the end effector and a lightweight improved YOLOv8-Pose neural network to detect cherry tomato bunches and picking points for robotic harvesting. The problem of occlusion and overlap is alleviated by merging RGB and depth images from the RGB-D sensor. To enhance detection robustness in complex backgrounds and reduce the complexity of the model, the Starblock module from StarNet and the coordinate attention mechanism are incorporated into the YOLOv8-Pose network, termed StarBL-YOLO, to improve the efficiency of feature extraction and reinforce spatial information. Additionally, we replaced the original OKS loss function with the L1 loss function for keypoint loss calculation, which improves the accuracy in picking points localization. The proposed method has been evaluated on a dataset with 843 cherry tomato RGB-D image pairs acquired by a harvesting robot at a commercial greenhouse farm. Experimental results demonstrate that the proposed StarBL-YOLO model achieves a 12% reduction in model parameters compared to the original YOLOv8-Pose while improving detection accuracy for cherry tomato bunches and picking points. Specifically, the model shows significant improvements across all metrics: for computational efficiency, model size (−11.60%) and GFLOPs (−7.23%); for pickable bunch detection, mAP50 (+4.4%) and mAP50-95 (+4.7%); for non-pickable bunch detection, mAP50 (+8.0%) and mAP50-95 (+6.2%); and for picking point detection, mAP50 (+4.3%), mAP50-95 (+4.6%), and RMSE (−23.98%). These results validate that StarBL-YOLO substantially enhances detection accuracy for cherry tomato bunches and picking points while improving computational efficiency, which is valuable for resource-constrained edge-computing deployment for harvesting robots.
Why it matches plant phenotyping methodsRGB-D画像と改良YOLOv8-Poseを用いて、収穫対象となるトマト房と摘採点を検出・位置推定する手法を開発し、実データセットで性能検証している。単なる収穫対象の位置検出に見えるが、房の可収穫性と摘採点という植物器官の状態・位置を抽出する技術的貢献が中心である。
abstractThis study proposes utilizing a low-cost off-the-shelf RGB-D sensor mounted on the end effector and a lightweight improved YOLOv8-Pose neural network to detect cherry tomato bunches and picking points for robotic harvesting.
GreenhouseRGB-D / ToFFruit2D/3D reconstructionSegmentationTrackingGrowth / development / phenology
Automatic fruit growth monitoring plays a vital role in advancing precision agriculture. Tracking the evolution of fruits over time is essential to monitor their development and optimize production. The ability to recognize fruits over periods of time, even with drastic scene changes, is a required capability of agricultural robots. This paper presents a system that allows long-term fruit tracking in 3D data. It generates instance-segmented 3D representations of plants at various growth stages over time, utilizing only consumer-grade RGB-D cameras installed on a mobile robot. Our approach first performs instance segmentation on each image in a sequence. Then, by exploiting geometric information and depth maps, we track the same instances throughout the sequence. We produce a 3D point cloud containing instances, exploiting odometry information and 3D semantic mapping. Once our robot performs a new recording at a different plant growth stage, it associates each fruit with the previously built 3D cloud and update the model. We validate the system in a real-world glasshouse environment in Bonn, Germany. Experimental results demonstrate that our system outperforms existing baselines even though it relies only on annotated images and operates at frame-rate, allowing the deployment on a real robot.
Why it matches plant phenotyping methodsRGB-D画像と3Dセマンティックマッピングによる果実の長期追跡・成長モニタリング手法が研究の中心であり、植物器官の成長状態を抽出するため、植物フェノタイピング手法として適格です。
abstractThis paper presents a system that allows long-term fruit tracking in 3D data.
PURPOSE: Efficient orchard management requires high-throughput phenotyping technologies to assist growers in crop monitoring and decision-making. This study presents Agrosense, an advanced artificial intelligence (AI) powered sensing system designed for real-time phenotypic data collection in orchards, addressing the limitations of traditional manual methods. METHODS: Agrosense integrates four RGB-D cameras with a Jetson Xavier microprocessor to collect high-resolution data and perform tree crop counting, canopy density classification, and tree height estimation. A citrus orchard served as a case study, where 337 trees were imaged to train and validate AI models. YOLOv8 was employed for object detection and classification tasks, while five methods were tested for estimating tree height. RESULTS: The YOLOv8 model achieved a mean average precision (mAP) of 0.977 for tree trunkdetection and 0.974 for canopy density classification. In field testing on 157 citrus trees, the system achieved 95% accuracy for tree trunk detection and 94% accuracy for canopy density classification, with only 11 misclassifications. The best-performing method for height estimation achieved a mean absolute percentage error (MAPE) of 8.53%. Agrosense completed phenotyping tasks in 398 s, a 515% speed improvement over manual methods (2,446 s). CONCLUSION: Agrosense effectively supports precision orchard management by automating key phenotyping tasks with high accuracy and efficiency. The system significantly reduces data collection time and improves consistency. Future work will focus on algorithm refinement and adaptation to other tree crops to broaden the system’s utility in precision agriculture.
Why it matches plant phenotyping methodsRGB-DカメラとAIを統合した果樹フェノタイピングシステムを開発・検証し、樹冠密度や樹高などの形質を定量化しているため、方法が研究の中心である。
abstractThis study presents Agrosense, an advanced artificial intelligence (AI) powered sensing system designed for real-time phenotypic data collection in orchards
Dwarf tomatoes, with high edible and ornamental value, require monitoring multiple growth parameters to balance yield and aesthetics. While deep learning has been widely applied in phenotype monitoring, most studies focus on individual growth parameters, overlooking intrinsic relationships. To simultaneously monitor multiple growth parameters across the entire growth stage and different cultivars, this study develops a multi-modal multi-task phenotype monitoring network for dwarf tomatoes (TomPhenoNet). The network model utilizes top-view RGB-D images to evaluate four key growth parameters: height, leaf area, fresh weight, and the number of red fruits. TomPhenoNet generates mask images, fruit detection features, and the number of detected fruits based on RGB images. By fusing RGB-D images, mask images, and fruit detection features, and introducing the cross-stitch network, the network predicts plant height, leaf area, and fresh weight. The predicted values are further used to generate the dynamic occlusion coefficient, adjusting the number of detected fruits to accurately predict the number of red fruits. Results reveal that TomPhenoNet achieves high prediction performances, with R² values of 0.828, 0.930, 0.945, and 0.881 for plant height, leaf area, fresh weight, and the number of red fruits, respectively. Ablation experiments show that the cross-stitch network and fruit detection features improve the prediction performances of growth parameters, with TomPhenoNet combining both modules performing best. Feature importance analysis indicates the network model captures plant growth characteristics and corrects the impact of leaf occlusion from the top view. This study promotes accurate tomato monitoring and provides data support for optimizing cultivation strategies.
Why it matches plant phenotyping methodsトマトの複数形質をRGB-D画像から推定するマルチモーダル・マルチタスク手法を開発し、性能評価とアブレーション実験まで行っており、表現型取得・推定法が研究の中心である。
abstractthis study develops a multi-modal multi-task phenotype monitoring network for dwarf tomatoes (TomPhenoNet).
The architecture of rice tillers plays a pivotal role in yield potential, yet conventional phenotyping methods have struggled to capture these intricate three-dimensional (3D) structures with high fidelity. In this study, a 3D model reconstruction method was developed specifically for rice tillers to overcome the challenges posed by their slender, feature-poor morphology in multi-view stereo-based 3D reconstruction. By applying strategically designed colorful reference markers, high-resolution 3D tiller models of 231 rice landraces were reconstructed. Accurate phenotyping was achieved by introducing ScaleCalculator, a software tool that integrated depth images from a depth camera to calibrate the physical sizes of the 3D models. The high efficiency of the 3D model-based phenotyping pipeline was demonstrated by extracting the following seven key agronomic traits: flag leaf length, panicle length, first internode length below the panicle, stem length, flag leaf angle, second leaf angle from the panicle, and third leaf angle. Genome-wide association studies (GWAS) performed with these 3D traits identified numerous candidate genes, nine of which had been previously confirmed in the literature. This work provides a 3D phenomics solution tailored for slender organs and offers novel insights into the genetic regulation of complex morphological traits in rice.
Why it matches plant phenotyping methodsイネ分げつの3D再構成とScaleCalculatorによるスケール校正を開発し、7つの形態形質を抽出するフェノタイピング手法が研究の中心であるため。
abstracta 3D model reconstruction method was developed specifically for rice tillers
Reproduction assets foundThe paper's 3D tiller models for 231 rice landraces are publicly deposited on Zenodo, and the authors' ScaleCalculator phenotyping source code is publicly available on GitHub, both explicitly stated in the Data Availability Statement. SNP genotype data are unpublished and excluded.Code · publicvelopment Co. LTD, and
Jiangsu Collaborative Innovation Center for Modern Crop Production.
Data Availability Statement: The 3D tiller models created in this study are available for research pur-
poses at https://zenodo.org/records/16080993 (accessed on 18 July 2025).The source code of ScaleCal-
culator is available on GitHub at https://github.com/ganlab/OSTRA/tree/master/ScaleCalculator
(accessed on 18 July 2025).
Acknowledgments: We thank Jianmin Wan for their valuable suggestions and Jiaqi Deng for their
technical help.
Conflicts of Interest: The authors declare that there are no conflicts of interest regarding the publica-
tion of this article.
References
1. Food and Agriculture OrganizatOpen asset ↗github · ganlab/OSTRApdf-raw-page:16 lines:1-50Plant phenotyping relevance match · UnverifiedOpenAlex · checked 13 Sept 2026
The acquisition of plant phenotypic traits is essential for selecting superior varieties, improving crop yield, and supporting precision agriculture and agricultural decision-making. Therefore, it plays a significant role in modern agriculture and plant science research. Traditional manual measurements of phenotypic traits are labor-intensive and inefficient. In contrast, combining 3D reconstruction technologies with autonomous vehicles enables more intuitive and efficient trait acquisition. This study proposes a 3D semantic reconstruction system based on an improved ORB-SLAM3 framework, which is mounted on an unmanned vehicle to acquire phenotypic traits in tomato cultivation scenarios. The vehicle is also equipped with the A * algorithm for autonomous navigation. To enhance the semantic representation of the point cloud map, we integrate the BiSeNetV2 network into the ORB-SLAM3 system as a semantic segmentation module. Furthermore, a two-stage filtering strategy is employed to remove outliers and improve the map accuracy, and OctoMap is adopted to store the point cloud data, significantly reducing the memory consumption. A spherical fitting method is applied to estimate the number of tomato fruits. The experimental results demonstrate that BiSeNetV2 achieves a mean intersection over union (mIoU) of 95.37% and a frame rate of 61.98 FPS on the tomato dataset, enabling real-time segmentation. The use of OctoMap reduces the memory consumption by an average of 96.70%. The relative errors when predicting the plant height, canopy width, and volume are 3.86%, 14.34%, and 27.14%, respectively, while the errors concerning the fruit count and fruit volume are 14.36% and 14.25%. Localization experiments on a field dataset show that the proposed system achieves a mean absolute trajectory error (mATE) of 0.16 m and a root mean square error (RMSE) of 0.21 m, indicating high localization accuracy. Therefore, the proposed system can accurately acquire the phenotypic traits of tomato plants, providing data support for precision agriculture and agricultural decision-making.
Why it matches plant phenotyping methodsRGB-D SLAM、自律走行、意味分割、点群処理を統合し、トマトの草丈・樹冠幅・体積・果実数・果実体積を推定するフェノタイピング手法の開発と検証が中心である。
abstractThis study proposes a 3D semantic reconstruction system based on an improved ORB-SLAM3 framework, which is mounted on an unmanned vehicle to acquire phenotypic traits in tomato cultivation scenarios.
In orchard automation, dense foliage during the canopy season severely occludes tree structures, minimizing visibility to various canopy parts such as trunks and branches, which limits the ability of a machine vision system. However, canopy structure is more open and visible during the dormant season when trees are defoliated. In this work, we present an information fusion framework that integrates multi-seasonal structural data to support robotic and automated crop load management during the entire growing season. The framework combines high-resolution RGB-D imagery from both dormant and canopy periods using YOLOv9-Seg for instance segmentation, Kinect Fusion for 3D reconstruction, and Fast Generalized Iterative Closest Point (Fast GICP) for model alignment. Segmentation outputs from YOLOv9-Seg were used to extract depth-informed masks, which enabled accurate 3D point cloud reconstruction via Kinect Fusion; these reconstructed models from each season were subsequently aligned using Fast GICP to achieve spatially coherent multi-season fusion. The YOLOv9-Seg model, trained on manually annotated images, achieved a mean squared error (MSE) of 0.0047 and segmentation mAP@50 scores up to 0.78 for trunks in dormant season dataset. Kinect Fusion enabled accurate reconstruction of tree geometry, validated with field measurements resulting in root mean square errors (RMSE) of 5.23 mm for trunk diameter, 4.50 mm for branch diameter, and 13.72 mm for branch spacing. Fast GICP achieved precise cross-seasonal registration with a minimum fitness score of 0.00197, allowing integrated, comprehensive tree structure modeling despite heavy occlusions during the growing season. This fused structural representation enables robotic systems to access otherwise obscured architectural information, improving the precision of pruning, thinning, and other automated orchard operations.
Why it matches plant phenotyping methods果樹の幹・枝の形態を3D再構成・季節間融合で推定する画像ベースの表現型取得手法を開発し、実測値で検証しているため、方法が中心である。
abstractIn this work, we present an information fusion framework that integrates multi-seasonal structural data to support robotic and automated crop load management during the entire growing season.
By capturing the intricate structural and spectral variations of the plant canopy, we can enhance our ability to model and predict dynamic parameters such as biomass with greater precision. This method not only preserves the plants for continuous monitoring but also provides a scalable and efficient alternative to traditional destructive techniques. The objective of this study was to examine the potential of using image-derived color and geometric plant features to output accurate predictions of three plant biomass accumulation parameters − leaf fresh weight, leaf dry weight, and leaf area for single plant monitoring. Top-view images of a hydroponic ‘Chicarita’ romaine lettuce (Lactuca sativa) crop captured with a color and depth sensor were used as the input of a multiple plants image processing workflow that extracted plant height, canopy morphometric, and color traits at an individual plant level. Two destructive harvest rounds were performed across the plant cycle to measure the observed values for each biomass response given by leaf fresh weight, leaf dry weight and leaf area from two crop cycles. The image-derived traits were used as potential predictors for a simple linear regression used as a baseline model and for two supervised machine learning models (random forest and least absolute shrinkage and selection operator or LASSO regression) to estimate each response. Using extracted depth information, vertical height per plant was estimated with a mean absolute error of 1.51 cm. Random Forest regression models yielded the most accurate predictions on a first harvest round for all three biomass parameters with R² values of 0.74, 0.80, and 0.67 and mean absolute percentage error (MAPE) of 11.77%, 10.16%, and 12.50%. LASSO regression outperformed the other models in a second harvest round with R² values of 0.72, 0.65, and 0.79 and MAPE of 7.79%, 7.76%, and 7.06% for leaf fresh weight, leaf dry weight, and leaf area, respectively. These results suggest that using a selection of canopy descriptors may improve the non-destructive biomass estimation along a lettuce crop cycle, enabling remote monitoring and real-time harvest projections.
Why it matches plant phenotyping methodsRGB-depth画像から植物形質を抽出し、機械学習でレタスのバイオマスを非破壊推定する手法が研究の中心であるため。
abstractimage processing workflow that extracted plant height, canopy morphometric, and color traits at an individual plant level
Field / plotLiDAR / point cloudRGB-D / ToFStem / branchWhole plant / canopy / plot / fieldClassificationMorphology / geometry measurementArchitecture / morphology / geometryGrowth / development / phenology
Effective assessment of tree maturity and agrochemical application requirements is important for optimizing resource use and sustainability in woody ornamental nurseries. This study integrates a ground-based red, green, blue – depth (RGB-D) camera and Light Detection and Ranging (LiDAR) system to measure key physiological parameters, trunk diameter and canopy volume, for maturity evaluation and precision spraying, respectively. Trunk diameter was calculated using a circle-fitting algorithm on point clouds at 0.15 meters (6 inches) above ground, derived from RGB-D pair segmented using Fast Segment Anything Model (FastSAM). Canopy volume was estimated by using a convex hull algorithm on processed point clouds through point cloud registration, ROI (region of interest) clipping, and denoising. Thirty-two trees were randomly selected in pairs from two plots (Plot-1 and Plot-2) with varying terrains for this experiment. The trunk diameter results in Plot-1 exhibited an average absolute error percentage of 0.23%, with an RMSE (root mean square error) of 0.03 meters and MAE (mean average error) of 0.02 meters, whereas Plot-2 showed an error percentage of 1.11%, with an RMSE of 0.08 meters and MAE of 0.07 meters. The trunk diameter was further analyzed for tree maturity analysis, revealing that Plot-1 had 10 mature trees while Plot-2 had only 5, indicating a more advanced growth stage in Plot-1. This classification was validated against manual assessments, showing 100% agreement across all 32 experimental trees, confirming the accuracy of the RGB-D system in determining tree maturity. Similarly, results for the canopy volume of Plot-1 indicated an average absolute error percentage of 10.99%, with RMSE and MAE values of 0.37 cubic meters and 0.33 cubic meters, respectively, while Plot-2 showed an error percentage of 13.01%, with an RMSE of 0.27 cubic meters and MAE of 0.24 cubic meters. These results demonstrate the feasibility and accuracy of integrating LiDAR and RGB-D technologies for efficient nursery management, supporting maturity assessment and precision agrochemical application as part of sustainable practices in ornamental horticulture.
Why it matches plant phenotyping methodsRGB-DとLiDARを用いた樹幹径・樹冠体積の測定手法を開発・検証し、成熟度判定を手動評価で検証しているため、植物表現型取得が中心的な研究です。
abstractThis study integrates a ground-based red, green, blue – depth (RGB-D) camera and Light Detection and Ranging (LiDAR) system to measure key physiological parameters, trunk diameter and canopy volume, for maturity evaluation and precision spraying, respectively.
Growth monitoring of tomato plants in large greenhouse environments is critical for quality and efficient production. Stem diameter and elongation are key phenotypic traits for plant growth monitoring. Traditional methods, however, rely on manual operations, which are time-consuming and labor-intensive and do not apply to large-scale greenhouses. Currently, automated image-based methods exemplified by three-dimensional (3D) point cloud technology are among the preferred solutions. Nevertheless, the occlusion of plant structures during the information acquisition process is challenging for practical applications. To address this challenge, this study proposes a novel method for plant stem occlusion inpainting using Deep Reinforcement Learning (DRL). Unlike most existing 3D reconstruction approaches that require depth data from multiple viewpoints, our solution captures 3D point cloud data from a single direction. The DRL model is applied to inpaint the incomplete stem for accurate stem reconstruction and phenotypic measurements. Specifically, our approach consists of two parts, structural completion and stem diameter completion. First, we extract the point cloud of incomplete stems from the RGB-D camera data. Second, we obtain the spatial structure of the stems by inpainting the 3D stem centerline with the DRL model. Finally, we add shape features (stem diameters) by inpainting the two edge lines of the stem occlusion part with the DRL model. For stem inpainted 3D point cloud data, we conducted validation experiments by measuring several commonly used stem phenotypic traits in tomato plants, including stem diameter, stem length, and stem inclination. The experimental results show that the Mean Absolute Percentage Error (MAPE) of the occluded main stem diameter is 9.7%, stem length is 5.7%, and tilt angle is 1%. For the occluded branch stem, the MAPE of stem diameter is 23.1%, stem length is 7.9%, and tilt angle is 1.5%. The accuracy of these measurements for occluded stems is acceptable compared to that obtained from 3D point clouds of unoccluded stems. This highlights the significant potential of using DRL to effectively inpaint occluded 3D point cloud data of plants.
Why it matches plant phenotyping methods植物茎の遮蔽部分を3D点群と深層強化学習で補完し、茎径・茎長・傾斜角という表現型形質を推定・検証する手法が研究の中心である。
abstractthis study proposes a novel method for plant stem occlusion inpainting using Deep Reinforcement Learning (DRL).
As the agricultural workforce declines and labor costs rise, robotic yield estimation has become increasingly important. While unmanned ground vehicles (UGVs) are commonly used for indoor farm monitoring, their deployment in greenhouses is often constrained by infrastructure limitations, sensor placement challenges, and operational inefficiencies. To address these issues, we develop a lightweight unmanned aerial vehicle (UAV) equipped with an RGB-D camera, a 3D LiDAR, and an IMU sensor. The UAV employs a LiDAR-inertial odometry algorithm for precise navigation in GNSS-denied environments and utilizes a 3D multi-object tracking algorithm to estimate the count and weight of cherry tomatoes. We evaluate the system using two dataset: one from a harvesting row and another from a growing row. In the harvesting-row dataset, the proposed system achieves 94.4\% counting accuracy and 87.5\% weight estimation accuracy within a 13.2-meter flight completed in 10.5 seconds. For the growing-row dataset, which consists of occluded unripened fruits, we qualitatively analyze tracking performance and highlight future research directions for improving perception in greenhouse with strong occlusions. Our findings demonstrate the potential of UAVs for efficient robotic yield estimation in commercial greenhouses.
Why it matches plant phenotyping methodsUAV上のRGB-D・LiDAR・追跡アルゴリズムにより、トマト果実数と重量を推定する手法を開発・評価しており、植物形質取得が研究の中心です。
abstractwe develop a lightweight unmanned aerial vehicle (UAV) equipped with an RGB-D camera, a 3D LiDAR, and an IMU sensor.
Grapevine phenotyping, that is the process of determining the physical properties (e.g., size, shape, and number) of grape bunches, provides valuable information for growth and health monitoring, yield estimation and efficient crop management in precision viticulture. Currently, grape bunch counting and sizing is done manually, which is labor intensive and often impractical for large-scale field applications. This paper describes a novel framework to automatically detect, count and estimate the volume/weight of grape bunches using RGB and depth data acquired in the field by a farmer robot. The proposed pipeline starts with the semantic segmentation of RGB images based on a pre-trained MANet architecture with EfficientnetB3 backbone to separate fruit from non-fruit regions. The segmented fruit mask is then projected onto the co-registered depth image to recover a depth mask, allowing for three-dimensional (3D) data association. After a pre-processing step to correct anomalies, such as corrupted and missing values, and to remove outliers, a depth gradient-based clustering algorithm is applied that detects individual grape bunch clusters. This enables the separation of adjacent and partially overlapping bunches. In addition, a method to reconstruct the whole 3D shape of a bunch is introduced, so as to provide an estimate of volume and weight. Experiments performed in a commercial vineyard in Italy are presented showing that, despite the low quality and high variability of the input images, the proposed approach is able to count grape bunch clusters with an average error of about 12% with respect to visual ground-truth and an average error less than 30% with respect to manual weight measurements. It is also shown that the processing framework can be applied to geo-referenced image sequences acquired by the farmer robot while traversing vineyard rows, thus providing an automated pipeline for the generation of high-resolution yield maps for precision viticulture applications.
Why it matches plant phenotyping methodsブドウ房の検出・計数・3D形状復元により体積・重量を推定する画像・深度ベースの表現型取得手法を開発し、精度検証も行っているため、方法が中心的である。
abstractThis paper describes a novel framework to automatically detect, count and estimate the volume/weight of grape bunches using RGB and depth data acquired in the field by a farmer robot.
Accurate and precise spraying in orchards is paramount for optimized agricultural practices, ensuring efficient pesticide utilization, minimized environmental impact, and enhanced crop yield by targeting specific areas with the right amount of treatment. The asymmetrical distribution of foliage and flowers in peach orchards poses a formidable challenge to achieving precise spray accuracy, impeding the uniform application of treatments and compromising the overall efficacy of pest and disease control measures. In response to the prevailing challenges in achieving accurate spray application caused by the asymmetrical distribution of foliage and flowers in peach orchards, this paper introduces a novel deep neural network to map the RGB image and corresponding depth to the density map of peach flowers or foliage. The model consists of components: (1) two backbones based on ResNet-50 that extract contextual features from the RGB image and depth features from depth data at multiple scales and levels; (2) an optimized depth-enhanced module that effectively fuses the distinct features extracted from the two input streams; and (3) a two-stage decoder that aggregates the high-level cross-modal features to regress the coarse density map and subsequently integrates it with the low-level cross-modal features for final density map prediction. To evaluate the performance of our model, we collected 493 frames (206,095 instances) of peach flowers and 475 frames (350,833 instances) of foliage from the peach orchards utilizing our sprayer prototype equipped with stereo cameras. The proposed method outperforms state-of-the-art models on our datasets, demonstrating the superiority and efficacy for encoding canopy characteristics in the form of flower and foliage density maps for blossom and cover sprays. It attains significant computational efficiency, exhibiting a frame rate of 20 FPS, and showcases exceptional accuracy with a WMAPE of 12.11% for peach flowers and a WMAPE of 13.37% for leaves.
Why it matches plant phenotyping methods桃の花・葉の密度という植物器官の形質をRGB-D画像から推定する手法を開発・評価しており、散布制御への応用を超えてフェノタイピング手法自体が中心です。
abstractthis paper introduces a novel deep neural network to map the RGB image and corresponding depth to the density map of peach flowers or foliage
Large-scale manual measurements of plant architectural traits in tomato growth are laborious and subjective, hindering deeper understanding of temporal variations in gene expression heterogeneity. This study develops a high-throughput approach for characterizing tomato architectural traits at different growth stages and mapping temporal broad-sense heritability using an unmanned ground vehicle-based plant phenotyping system. The SegFormer with fusion of multispectral and depth imaging modalities was employed to semantically segment plant organs from the registered RGB-D and multispectral images. Organ point clouds were then generated and clustered into instances. Finally, six key architectural traits, including fruit spacing (FS), inflorescence height (IH), stem thickness (ST), leaf spacing (LS), total leaf area (TLA), and leaf inclination angle (LIA) were extracted and the temporal broad-sense heritability folds were plotted. The root mean square errors (RMSEs) of the estimated FS, IH, ST, and LS were 0.014, 0.043, 0.003, and 0.015 m, respectively. The visualizations of the estimated TLA and LIA matched the actual growth trends. The broad-sense heritability of the extracted traits exhibited different trends across the growth stages: (i) ST, IH, and FS had a gradually increased broad-sense heritability over time, (ii) LS and LIA had a decreasing trend, and (iii) TLA showed fluctuations (i.e. an M-shaped pattern) of the broad-sense heritability throughout the growth period. The developed system and analytical approach are promising tools for accurate and rapid characterization of spatiotemporal changes of tomato plant architecture in controlled environments, laying the foundation for efficient crop breeding and precision production management in the future.
Why it matches plant phenotyping methods植物形態形質を取得するUGV型マルチモーダル画像フェノタイピングシステムと解析手法の開発・定量評価が研究の中心であり、誤差検証も行っているため。
abstractThis study develops a high-throughput approach for characterizing tomato architectural traits at different growth stages and mapping temporal broad-sense heritability using an unmanned ground vehicle-based plant phenotyping system.
Reproduction assets foundThe paper's Data availability statement explicitly deposits the authors' trait-extraction pipeline code and example data on a public GitHub repository, matching the allowed URL.Code · publicThe pipeline code and example data related to this project are available as open source on GitHub ( https://github.com/DigBigPigForU/Tomato-architectural-trait-extraction ).Open asset ↗Tomato-architectural-trait-extractionlines:822-958Plant phenotyping relevance match · UnverifiedOpenAlex · checked 13 Sept 2026
Orchard yield estimation is one of the key indicators of precision agriculture. The traditional random sampling yield estimation method has strict requirements for the laborer experience and scale of orchards. Intelligent orchard management enables growers to use resources more effectively and make wiser decisions to optimize orchard inputs. This study proposes a banana bunch counting and yield estimation method based on the DeepSORT tracking algorithm. This method involves obtaining RGB-D images and calculating the weight of an individual bunch of bananas, which was promoted in our previous work. Building on this, the DeepSORT was used to solve the repeated counting based on the Hungarian algorithm and Kalman filtering. Three constraints were set to improve the statistical accuracy, and a yield estimation system was designed for orchard management monitoring. This system provides managers with bunch weight predictions and statistical plant information to achieve real-time yield estimations for banana orchards. The experimental results showed that the accuracy of the yield estimations reached 97.25% and that banana bunch counting had a success rate of 96.82%. This demonstrates that the effective integration of RGB-D technology and the DeepSORT algorithm can be successfully applied to the intelligent management and harvesting of banana orchards.
Why it matches plant phenotyping methodsRGB-D画像とDeepSORTによるバナナ房の計数・重量推定という、植物の収量形質を抽出する画像ベース手法が研究の中心であり、精度評価も実施しているため。
abstractThis study proposes a banana bunch counting and yield estimation method based on the DeepSORT tracking algorithm.
To address the low estimation accuracy of deep learning-based crop yield image recognition methods under untrained shooting distances, this study proposes a shooting distance adaptive crop yield estimation method by fusing RGB and depth image information through multi-modal data fusion. Taking strawberry fruit fresh weight as an example, RGB and depth image data of 348 strawberries were collected at nine heights ranging from 70 to 115 cm. First, based on RGB images and shooting height information, a single-modal crop yield estimation model was developed by training a convolutional neural network (CNN) after cropping strawberry fruit images using the relative area conversion method. Second, the height information was expanded into a data matrix matching the RGB image dimensions, and multi-modal fusion models were investigated through input-layer and output-layer fusion strategies. Finally, two additional approaches were explored: direct fusion of RGB and depth images, and extraction of average shooting height from depth images for estimation. The models were tested at two untrained heights (80 cm and 100 cm). Results showed that when using only RGB images and height information, the relative area conversion method achieved the highest accuracy, with R2 values of 0.9212 and 0.9304, normalized root mean square error (NRMSE) of 0.0866 and 0.0814, and mean absolute percentage error (MAPE) of 0.0696 and 0.0660 at the two untrained heights. By further incorporating depth data, the highest accuracy was achieved through input-layer fusion of RGB images with extracted average height from depth images, improving R2 to 0.9475 and 0.9384, reducing NRMSE to 0.0707 and 0.0766, and lowering MAPE to 0.0591 and 0.0610. Validation using a developed shooting distance adaptive crop yield estimation platform at two random heights yielded MAPE values of 0.0813 and 0.0593. This model enables adaptive crop yield estimation across varying shooting distances, significantly enhancing accuracy under untrained conditions.
Why it matches plant phenotyping methodsRGB・深度画像とCNNを用いてイチゴ果実重量（収量）を推定する撮影距離適応型の方法とプラットフォームを開発・検証しており、表現型取得・推定が研究の中心である。
abstractthis study proposes a shooting distance adaptive crop yield estimation method by fusing RGB and depth image information through multi-modal data fusion.
• Achieved high detection accuracy of apple fruitlets in complex orchard environments with rapid phenological changes. • Provided a dataset of videos and RGB-D images, featuring annotated apple fruitlets and manual caliper measurements during early development. • Developed a workflow for rapid in-field monitoring of flower corymbs and fruitlet sizing, validated through experimental trials. Current research in apple-growing focuses on collecting extensive biometric data to better understand physiological processes, improve orchard productivity and predict yields. In this context, fruit thinning has emerged as a key horticultural practice to enhance fruit size and quality while preventing alternate bearing. Despite the growing role of plant imaging technologies in agronomic management, fruitlet sizing remains challenging, particularly in early phenological stages. To address this challenge, we developed an RGB-D-based vision pipeline that combines YOLO models with depth information and relies on the statistical analysis of frame series to detect and cluster fruitlets into flower corymbs, providing both fruitlet counting and diameter estimates for each video acquisition. After obtaining an AP@0.5 and AP@[0.5:0.95] of respectively 0.894 and 0.77 in fruitlet detection, along with a precision of 0.881 and a recall of 0.846, our approach efficiently processed video frames, extracting the most reliable data for each labeled cluster. While the comparison of true positive estimates with calibrated caliper measurements showed a mean RMSE of 1.05 mm, challenges remain in achieving the correct fruitlet count, with a mean counting error of 0.63 fruitlets per video. Additionally, the proposed workflow retrieved the exact number of fruitlets as the ground truth in 56.4% of the videos, increasing to 75% when excluding those videos where the correct fruitlet count was never detected in any frame by the YOLO model. Despite these limitations, our results are promising, proposing a potential data acquisition tool without compromising the reliability of traditional practices. This approach could pave the way for future applications, including the evaluation of plant growth regulator trials and the development of predictive models for yield and productivity optimization.
Why it matches plant phenotyping methodsRGB-D画像と深度情報、YOLO、動画フレーム統計を組み合わせ、リンゴ果実の検出・計数・直径推定を行うワークフローを開発し、ノギス測定で検証しているため、表現型取得手法が中心である。
abstractProvided a dataset of videos and RGB-D images, featuring annotated apple fruitlets and manual caliper measurements during early development.
Accurate data acquisition of crop morphological parameters is crucial for effective greenhouse management decision-making and remote sensing technologies are increasingly being applied to automate the data collection process. This research utilised an RGB-D based computer vision method to investigate the correlation between the computer vision features and the lettuce morphological parameters, including leaf area, plant height, diameter, and fresh weight. A dataset of lettuce containing over 300 RGB images and depth images of the 3rd Autonomous Greenhouse Challenge was used, and Random Forest, XGBoost and linear regression models were applied in the prediction. The best NRMSE values for diameter, dry matter content, dry weight, fresh weight, height, and leaf area are 0.08, 0.08, 0.07, 0.07, 0.08, and 0.07, which showed a promising accuracy compared to similar studies. This research demonstrates a novel approach to non-destructively estimate greenhouse leafy vegetable morphological parameters.
Why it matches plant phenotyping methodsRGB-D画像と機械学習によりレタスの形態形質を非破壊推定する手法が研究の中心であり、植物フェノタイピング方法に該当する。
abstractThis research utilised an RGB-D based computer vision method to investigate the correlation between the computer vision features and the lettuce morphological parameters
Dwarf tomatoes, with high edible and ornamental value, require monitoring multiple growth parameters to balance yield and aesthetics. While deep learning has been widely applied in phenotype monitoring, most studies focus on individual growth parameters, overlooking intrinsic relationships. To simultaneously monitor multiple growth parameters across the entire growth stage and different cultivars, this study develops a multi-modal multi-task phenotype monitoring network for dwarf tomatoes (TomPhenoNet). The network model utilizes top-view RGB-D images to evaluate four key growth parameters: height, leaf area, fresh weight, and the number of red fruits. TomPhenoNet generates mask images, fruit detection features, and the number of detected fruits based on RGB images. By fusing RGB-D images, mask images, and fruit detection features, and introducing the cross-stitch network, the network predicts plant height, leaf area, and fresh weight. The predicted values are further used to generate the dynamic occlusion coefficient, adjusting the number of detected fruits to accurately predict the number of red fruits. Results reveal that TomPhenoNet achieves high prediction performances, with R 2 values of 0.828, 0.930, 0.945, and 0.881 for plant height, leaf area, fresh weight, and the number of red fruits, respectively. Ablation experiments show that the cross-stitch network and fruit detection features improve the prediction performances of growth parameters, with TomPhenoNet combining both modules performing best. Feature importance analysis indicates the network model captures plant growth characteristics and corrects the impact of leaf occlusion from the top view. This study promotes accurate tomato monitoring and provides data support for optimizing cultivation strategies.
Why it matches plant phenotyping methodsトマトの複数形質をRGB-D画像から推定するマルチモーダル・マルチタスク手法を開発し、性能評価とアブレーション実験を行っており、表現型取得・推定が研究の中心である。
abstractthis study develops a multi-modal multi-task phenotype monitoring network for dwarf tomatoes (TomPhenoNet).
Strawberry grading by picking robots can eliminate the manual classification, reducing labor costs and minimizing the damage to the fruit. Strawberry size or weight is a key factor in grading, with accurate weight estimation being crucial for proper classification. In this paper, we collected 1521 sets of strawberry RGB-D images using a depth camera and manually measured the weight and size of the strawberries to construct a training dataset for the strawberry weight regression model. To address the issue of incomplete depth images caused by environmental interference with depth cameras, this study proposes a multimodal point cloud completion method specifically designed for symmetrical objects, leveraging RGB images to guide the completion of depth images in the same scene. The method follows a process of locating strawberry pixel regions, calculating centroid coordinates, determining the symmetry axis via PCA, and completing the depth image. Based on this approach, a multimodal fusion regression model for strawberry weight estimation, named MMF-Net, is developed. The model uses the completed point cloud and RGB image as inputs, and extracts features from the RGB image and point cloud by EfficientNet and PointNet, respectively. These features are then integrated at the feature level through gradient blending, realizing the combination of the strengths of both modalities. Using the Percent Correct Weight (PCW) metric as the evaluation standard, this study compares the performance of four traditional machine learning methods, Support Vector Regression (SVR), Multilayer Perceptron (MLP), Linear Regression, and Random Forest Regression, with four point cloud-based deep learning models, PointNet, PointNet++, PointMLP, and Point Cloud Transformer, as well as an image-based deep learning model, EfficientNet and ResNet, on single-modal datasets. The results indicate that among traditional machine learning methods, the SVR model achieved the best performance with an accuracy of 77.7% (PCW@0.2). Among deep learning methods, the image-based EfficientNet model obtained the highest accuracy, reaching 85% (PCW@0.2), while the PointNet + + model demonstrated the best performance among point cloud-based models, with an accuracy of 54.3% (PCW@0.2). The proposed multimodal fusion model, MMF-Net, achieved an accuracy of 87.66% (PCW@0.2), significantly outperforming both traditional machine learning methods and single-modal deep learning models in terms of precision.
Why it matches plant phenotyping methodsイチゴのRGB-D画像から重量・サイズを推定する画像／点群解析手法を開発し、複数モデルと比較評価しており、植物形質取得が中心である。
abstractthis study proposes a multimodal point cloud completion method specifically designed for symmetrical objects
In facility tomato cultivation, analyzing and managing crop phenotypic traits across different growth stages is vital for improving quality and yield. Traditional manual methods are time-consuming and labor-intensive, particularly in densely cultivated environments. This study presents an automated pipeline for extracting key phenotypic parameters, including plant height, stem diameter, and fruit clusters, through three-dimensional (3D) model construction, organ segmentation, and phenotypic analysis. Point cloud data were captured at the seedling, flowering, and fruiting stages using Kinect v2 cameras and registered with the FPFH-ICP algorithm to mitigate occlusion-induced data loss. To better extract phenotypic parameters such as stem diameter and fruit clusters, constructed four multi-feature point clouds datasets, incorporating geometric coordinates, normal vectors, and color information (XYZ, XYZ-RGB, XYZ-Normal, and XYZ-Normal-RGB). RGB data highlighted regions with distinct color contrasts, such as fruits and leaves, while normal vectors captured surface details, enhancing the description of fine structures. To improve segmentation performance, we introduced CAFPoint, a modified PointNet++ model incorporating a multibranch structure and a CrossAttentionFusion (CAF) module. This structure can extract and fuse XYZ, normal and color information in a targeted manner. When applied to XYZ-Normal-RGB data, CAFPoint achieved an accuracy of 0.959 and a mean intersection over union (mIoU) of 0.863, demonstrating its capacity to enhance boundary delineation and feature representation. Phenotypic parameters derived from the segmentation, including stem diameter, fruit cluster count, and plant height, showed strong correlations with manual measurements (R² > 0.79 for stem diameter and fruit cluster count; R² > 0.92 for plant height). The proposed method enables efficient and accurate phenotypic trait extraction for tomatoes at various growth stages in greenhouse facilities, offering a valuable reference for automated trait analysis in controlled agricultural environments.
Why it matches plant phenotyping methods3D点群取得、器官セグメンテーション、深層学習モデルを統合し、トマトの草丈・茎径・果房数を自動抽出して手測定と検証しており、表現型取得手法が研究の中心である。
abstractThis study presents an automated pipeline for extracting key phenotypic parameters, including plant height, stem diameter, and fruit clusters, through three-dimensional (3D) model construction, organ segmentation, and phenotypic analysis.
Monitoring flowers over time is essential for precision robotic pollination in agriculture. To accomplish this, a continuous spatial-temporal observation of plant growth can be done using stationary RGB-D cameras. However, image registration becomes a serious challenge due to changes in the visual appearance of the plant caused by the pollination process and occlusions from growth and camera angles. Plants flower in a manner that produces distinct clusters on branches. This paper presents a method for matching flower clusters using descriptors generated from RGB-D data and considers allowing for spatial uncertainty within the cluster. The proposed approach leverages the Unscented Transform to efficiently estimate plant descriptor uncertainty tolerances, enabling a robust image-registration process despite temporal changes. The Unscented Transform is used to handle the nonlinear transformations by propagating the uncertainty of flower positions to determine the variations in the descriptor domain. A Monte Carlo simulation is used to validate the Unscented Transform results, confirming our method's effectiveness for flower cluster matching. Therefore, it can facilitate improved robotics pollination in dynamic environments.
Why it matches plant phenotyping methodsRGB-D画像から花房を時系列追跡・マッチングする画像登録手法が研究の中心で、植物成長の観測に直接用いられるため、植物フェノタイピング手法として含める。
abstractThis paper presents a method for matching flower clusters using descriptors generated from RGB-D data and considers allowing for spatial uncertainty within the cluster.
Crop yield estimation is a relevant problem in agriculture, because an accurate yield estimate can support farmers' decisions on harvesting or precision intervention. Robots can help to automate this process. To do so, they need to be able to perceive the surrounding environment to identify target objects such as trees and plants. In this paper, we introduce a novel approach to address the problem of hierarchical panoptic segmentation of apple orchards on 3D data from different sensors. Our approach is able to simultaneously provide semantic segmentation, instance segmentation of trunks and fruits, and instance segmentation of trees (a trunk with its fruits). This allows us to identify relevant information such as individual plants, fruits, and trunks, and capture the relationship among them, such as precisely estimate the number of fruits associated to each tree in an orchard. To efficiently evaluate our approach for hierarchical panoptic segmentation, we provide a dataset designed specifically for this task. Our dataset is recorded in Bonn, Germany, in a real apple orchard with a variety of sensors, spanning from a terrestrial laser scanner to a RGB-D camera mounted on different robots platforms. The experiments show that our approach surpasses state-of-the-art approaches in 3D panoptic segmentation in the agricultural domain, while also providing full hierarchical panoptic segmentation. Our dataset is publicly available at https://www.ipb.uni-bonn.de/data/hops/. The open-source implementation of our approach is available at https://github.com/PRBonn/hapt3D.
Why it matches plant phenotyping methodsリンゴ樹・果実・幹を3Dセグメンテーションし、樹ごとの果実数を推定する手法と専用データセットを中心に開発・評価しており、植物の器官形態・収量関連形質の取得に該当する。
abstractwe introduce a novel approach to address the problem of hierarchical panoptic segmentation of apple orchards on 3D data from different sensors.
Reproduction assets foundThe paper introduces the HOPS dataset of annotated 3D apple orchard point clouds (TLS, UAV, UGV, SfM) for hierarchical panoptic segmentation, publicly available at the authors' IPB Bonn page, and releases the open-source implementation (hapt3D) on GitHub. Both are paper-specific, public, and actionable.Code · publicThe open-source implementation of our approach is available at https://github.com/PRBonn/hapt3D .Open asset ↗PRBonn/hapt3Dlines:1-59Plant phenotyping relevance match · UnverifiedOpenAlex · arXiv · checked 15 Sept 2026
Estimation of a single leaf area can be a measure of crop growth and a phenotypic trait to breed new varieties. It has also been used to measure leaf area index and total leaf area. Some studies have used hand-held cameras, image processing 3D reconstruction and unsupervised learning-based methods to estimate the leaf area in plant images. Deep learning works well for object detection and segmentation tasks; however, direct area estimation of objects has not been explored. This work investigates deep learning-based leaf area estimation, for RGBD images taken using a mobile camera setup in real-world scenarios. A dataset for attached leaves captured with a top angle view and a dataset for detached single leaves were collected for model development and testing. First, image processing-based area estimation was tested on manually segmented leaves. Then a Mask R-CNN-based model was investigated, and modified to accept RGBD images and to estimate the leaf area. The detached-leaf data set was then mixed with the attached-leaf plant data set to estimate the single leaf area for plant images, and another network design with two backbones was proposed: one for segmentation and the other for area estimation. Instead of trying all possibilities or random values, an agile approach was used in hyperparameter tuning. The final model was cross-validated with 5-folds and tested with two unseen datasets: detached and attached leaves. The F1 score with 90% IoA for segmentation result on unseen detached-leaf data was 1.0, while R-squared of area estimation was 0.81. For unseen plant data segmentation, the F1 score with 90% IoA was 0.59, while the R-squared score was 0.57. The research suggests using attached leaves with ground truth area to improve the results.
Why it matches plant phenotyping methodsRGBD画像と深層学習を用いて葉面積という植物形質を推定する手法を開発し、交差検証と未知データで性能評価しており、フェノタイピング手法が中心である。
abstractThis work investigates deep learning-based leaf area estimation, for RGBD images taken using a mobile camera setup in real-world scenarios.
Growing consumer demand for high-quality strawberries has highlighted the need for accurate, efficient, and non-destructive methods to assess key postharvest quality traits, such as weight, size uniformity, and quantity. This study proposes a multi-objective learning algorithm that leverages RGB-D multimodal information to estimate these quality metrics. The algorithm develops a fusion expert network architecture that maximizes the use of multimodal features while preserving the distinct details of each modality. Additionally, a novel Heritable Loss function is implemented to reduce redundancy and enhance model performance. Experimental results show that the coefficient of determination (R²) values for weight, size uniformity and number are 0.94, 0.90 and 0.95 respectively. Ablation studies demonstrate the advantage of the architecture in multimodal, multi-task prediction accuracy. Compared to single-modality models, non-fusion branch networks, and attention-enhanced fusion models, our approach achieves enhanced performance across multi-task learning scenarios, providing more precise data for trait assessment and precision strawberry applications.
Why it matches plant phenotyping methodsRGB-D画像を用いてイチゴ果実の重量・サイズ均一性・個数を推定する融合ネットワークの開発と性能評価が研究の中心であり、植物器官の形質取得手法に該当する。
abstractThis study proposes a multi-objective learning algorithm that leverages RGB-D multimodal information to estimate these quality metrics.
Achieving global sustainable agriculture requires farmers worldwide to adopt smart agricultural technologies, such as autonomous ground robots. However, most ground robots are either task- or crop-specific and expensive for small-scale farmers and smallholders. Therefore, there is a need for cost-effective robotic platforms that are modular by design and can be easily adapted to varying tasks and crops. This paper describes the hardware design of a unique, low-cost multiaxial modular agricultural robot (ModagRobot), and its field evaluation for soybean phenotyping. The ModagRobot’s chassis was designed without any welded components, making it easy to adjust trackwidth, height, ground clearance, and length. For this experiment, the ModagRobot was equipped with an RGB-Depth (RGB-D) sensor and adapted to safely navigate over soybean rows to collect RGB-D images for estimating soybean phenotypic traits. RGB images were processed using the Excess Green Index to estimate the percent canopy ground coverage area. 3D point clouds generated from RGB-D images were used to estimate canopy height (CH) and the 3D Profile Index of sample plots using linear regression. Aboveground biomass (AGB) was estimated using extracted phenotypic traits. Results showed an R2, RMSE, and RRMSE of 0.786, 0.0181 m, and 2.47%, respectively, between estimated CH and measured CH. AGB estimated using all extracted traits showed an R2, RMSE, and RRMSE of 0.59, 0.0742 kg/m2, and 8.05%, respectively, compared to the measured AGB. The results demonstrate the effectiveness of the ModagRobot for in-row crop phenotyping.
Why it matches plant phenotyping methods低コスト移動ロボット、RGB-D撮像、画像・点群解析によるダイズ形質推定を開発・評価しており、植物フェノタイピング手法が中心である。
abstractThis paper describes the hardware design of a unique, low-cost multiaxial modular agricultural robot (ModagRobot), and its field evaluation for soybean phenotyping.
Accurate tomato yield estimation and ripeness monitoring are critical for optimizing greenhouse management. While manual counting remains labor-intensive and error-prone, this study introduces a novel vision-based framework for automated tomato counting in standardized greenhouse environments. The proposed method integrates YOLOv8-based detection, depth filtering, and an inter-frame prediction algorithm to address key challenges such as background interference, occlusion, and double-counting. Our approach achieves 97.09% accuracy in tomato cluster detection, with mature and immature single-fruit recognition accuracies of 92.03% and 91.79%, respectively. The multi-target tracking algorithm demonstrates a MOTA (Multiple Object Tracking Accuracy) of 0.954, outperforming conventional methods like YOLOv8+DeepSORT. By fusing odometry data from an inspection robot, this lightweight solution enables real-time yield estimation and maturity classification, offering practical value for precision agriculture.
Why it matches plant phenotyping methods温室ロボット向けの画像解析フレームワークを中心に、トマト果実の計数、成熟度分類、収量推定という植物器官・状態の定量手法を開発・評価しているため。
abstractthis study introduces a novel vision-based framework for automated tomato counting in standardized greenhouse environments.
This study systematically performed an extensive real-world evaluation of the performances of all configurations of YOLOv8, YOLOv9, YOLOv10, YOLO11( or YOLOv11), and YOLOv12 object detection algorithms in terms of precision, recall, mean Average Precision at 50\% Intersection over Union (mAP@50), and computational speeds including pre-processing, inference, and post-processing times immature green apple (or fruitlet) detection in commercial orchards. Additionally, this research performed and validated in-field counting of the fruitlets using an iPhone and machine vision sensors. Among the configurations, YOLOv12l recorded the highest recall rate at 0.90, compared to all other configurations of YOLO models. Likewise, YOLOv10x achieved the highest precision score of 0.908, while YOLOv9 Gelan-c attained a precision of 0.903. Analysis of mAP@0.50 revealed that YOLOv9 Gelan-base and YOLOv9 Gelan-e reached peak scores of 0.935, with YOLO11s and YOLOv12l following closely at 0.933 and 0.931, respectively. For counting validation using images captured with an iPhone 14 Pro, the YOLO11n configuration demonstrated outstanding accuracy, recording RMSE values of 4.51 for Honeycrisp, 4.59 for Cosmic Crisp, 4.83 for Scilate, and 4.96 for Scifresh; corresponding MAE values were 4.07, 3.98, 7.73, and 3.85. Similar performance trends were observed with RGB-D sensor data. Moreover, sensor-specific training on Intel Realsense data significantly enhanced model performance. YOLOv11n achieved highest inference speed of 2.4 ms, outperforming YOLOv8n (4.1 ms), YOLOv9 Gelan-s (11.5 ms), YOLOv10n (5.5 ms), and YOLOv12n (4.6 ms), underscoring its suitability for real-time object detection applications.
Why it matches plant phenotyping methods果実数という植物器官形質の画像ベース推定を対象に、複数YOLOモデルの性能比較とiPhone・RGB-Dセンサーによる圃場カウント検証を中心的に行っているため。
abstractThis study systematically performed an extensive real-world evaluation of the performances of all configurations of YOLOv8, YOLOv9, YOLOv10, YOLO11( or YOLOv11), and YOLOv12 object detection algorithms
Abstract Monitoring plant growth is crucial for effective crop management, and using color and depth (RGBD) cameras to model lettuce has emerged as one of the most convenient and non-invasive methods. In recent years, deep learning techniques, particularly neural networks, have become popular for estimating lettuce fresh weight. However, these models are typically specific to particular datasets, lack domain adaptation, and are often limited by the availability of open-access datasets. In this study, we propose a method based on plant geometric features for estimating the rosette structure and volume of lettuce. This new approach was compared to existing methods that reconstruct surfaces from point clouds, such as Ball Pivoting and Alpha Shapes. The proposed method creates a tight hull around the plant's point cloud, preserving high detail of the rosette structure while filling in surface holes in areas not visible to 3D cameras. Using a linear regression model, we estimated fresh weight for this dataset, achieving a root mean square error (RMSE) of 18.2 g when using only the estimated plant volume, and 17.3 g when both volume and geometric features were included. Additionally, we introduced new geometric features that characterize leaf density, which could be useful for breeding applications. A dataset of 402 point clouds of lettuce plants, captured before harvest, was compiled using one top-down and three side-view 3D cameras.
Why it matches plant phenotyping methodsRGB-D画像からレタスの構造・体積・葉密度を抽出し、生体重推定を検証する手法開発が研究の中心であり、データセットも構築している。
abstractIn this study, we propose a method based on plant geometric features for estimating the rosette structure and volume of lettuce.
Reproduction assets foundThe paper's own lettuce 3D point cloud dataset (Pii, 402 point clouds with fresh weight references) is deposited on Zenodo, and the vacuum-package surface reconstruction code plus data processing scripts are publicly available on the authors' GitHub repository. Both are paper-specific, public, and actionable.Dataset · publicData used in this study and developed models are available on Zenodo storage service https://zenodo.org/records/8410252 .Open asset ↗Zenodo · 8410252lines:158-220Code · publicThe code used at this study is available at https://github.com/VicB18/LettuceFW (accessed on 1 November 2024).Open asset ↗GitHub · VicB18/LettuceFWlines:158-220Code · publicThe code for the vacuum package method, along with the data processing scripts used in this study, are available in the Supplementary Information.Open asset ↗lines:98-114Plant phenotyping relevance match · UnverifiedOpenAlex · checked 13 Sept 2026
Green asparagus grows in clusters, which can cause overlaps with weeds and immature stems, making it difficult to identify suitable harvest targets and cutting points. Extracting precise stem details in complex spatial arrangements is a challenge. This paper explored the YS3AM (Yolo-SAM-3D-Adaptive-Modeling) method for detecting green asparagus and performing 3D adaptive-section modeling using a depth camera, which could benefit harvesting path planning for selective harvesting robots. Firstly, the model was developed and deployed to extract bounding boxes for individual asparagus stems within clusters. Secondly, the stems inside these bounding boxes were segmented, and binary masks were generated. Thirdly, high-quality depth images were obtained through pixel block completion. Finally, a novel 3D reconstruction method, based on adaptive section modeling and combining the mask and depth data, is proposed. And an evaluation method is introduced to assess modeling accuracy. Experimental validation showed high-performance detection (1095 field images demonstrated, Precision: 98.75%, Recall: 95.46%, F1: 0.97) and robust 3D modeling (103 asparagus stems, average RMSE: length 0.74, depth: 1.105) under varying illumination conditions. The system achieved 22 ms per stem processing speed, enabling real-time operation. The results demonstrated that the 3D model accurately represents the spatial distribution of clustered green asparagus, enabling precise identification of harvest targets and cutting points. This model provided essential spatial pathways for end-effector path planning, thereby fulfilling the operational requirements for efficient green asparagus harvesting robots.
Why it matches plant phenotyping methods深度画像・マスクに基づくアスパラガス茎の3D再構成法を開発し、茎の長さ・深さと空間分布を定量化して精度検証している。収穫対象検出だけでなく、植物器官形態の抽出が中心である。
abstracta novel 3D reconstruction method, based on adaptive section modeling and combining the mask and depth data, is proposed.
In the era of machine learning-driven plant imaging, the production of annotated datasets is a very important contribution. In this data paper, a unique annotated dataset of seedling emergence kinetics is proposed. It is composed of almost 70,000 RGB-depth frames and more than 700,000 plant annotations. The dataset is shown valuable for training deep learning models and performing high-throughput phenotyping by imaging. The ability of such models to generalize to several species and outperform the state-of-the-art owing to the delivered dataset is demonstrated. We also discuss how this dataset raises new questions in plant phenotyping.
Why it matches plant phenotyping methods植物の出芽速度を対象とする大規模RGB深度画像・アノテーションデータセットを提供し、深層学習および高スループット表現型解析への利用性を実証しており、表現型取得基盤が中心である。
abstracta unique annotated dataset of seedling emergence kinetics is proposed
Reproduction assets foundThis is a data paper whose core contribution is a public annotated RGB-depth seedling dataset (~70,000 frames, >700,000 annotations) deposited in DATA INRAE with DOI 10.57745/AMFJTK, explicitly stated as publicly accessible. Other allowed URLs (license, Intel datasheet, Jülich record) are not paper-specific assets.Dataset · publicSynthesis of the full time-lapse and RGB-Depth full frame quantity per species
Species
Pots time-lapse
Labelled pots time-lapse
RGB-depth full frame
Rapeseed
1 760
336
15 218
Tomatoes
1 960
480
33 283
Beans
2 320
400
21 445
Total
6 040
1 216
69 946
The dataset is publicly accessible in the DATA INRAE repository, DOI: https://doi.org/10.57745/AMFJTK . The file tree structure is illustrated in Fig. 4 . The dataset is organized into 11 compressed .zip files, each corresponding to a distinct trial. Within these files, images are sorted chronologically by acquisition start date, then by camera, and stored in .png format within dedicated color and depth folders. Labels are alsoOpen asset ↗DATA INRAE · 10.57745/AMFJTKlines:105-195Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Published1 Feb 2025Computers and Electronics in Agriculture.
Shortage of labor in fruit crop production has become a significant challenge in recent years. Therefore, mechanized and automated machines have emerged as promising alternatives to labor-intensive orchard operations such as harvesting, pruning, and thinning. One of the key aspects of the automated machines in accomplishing these tasks is their ability to identify tree canopy parts such as trunk and branches and estimate their geometric and topological parameters such as branch diameter, branch length, branch angles, and spacing between branches. By utilizing geometric parameters such as branch diameter, length, and orientation, researchers then can develop automated pruning and thinning systems that make more effective decisions to achieve optimal fruit yield and quality by accurately estimating the desired crop-load. In this study, we propose a machine vision system for estimating one of the canopy parameters in apple orchards: branch diameter. This parameter was used to estimate the optimal number of fruit that individual branches could bear in a commercial orchard, which provides a basis for robotic pruning, flower thinning, and fruitlet thinning so that desired fruit yield and quality could be achieved. Utilizing color and depth information collected with an RGB-D sensor (Azure Kinect DK, Microsoft, Redmond, WA), a YOLOv8-based instance segmentation technique was developed to identify trunks and branches of apple trees in the dormant season. We then applied a Principal Component Analysis (PCA) technique to estimate branch orientation, which was subsequently utilized to estimate branch diameter. The estimated branch diameter was used to calculate the Limb Cross Sectional Area (LCSA), which was then used to estimate optimal crop-load, as a larger LCSA indicates a higher potential fruit-bearing capacity of the branch. With this approach, Root Mean Squared Error (RMSE) for branch diameter estimation was calculated to be 2.06 mm (relative RMSE 10.82%) and the same for crop-load estimation (Number of fruits per branch) to be 3.93 (relative RMSE 22.25%). Our study demonstrated a promising workflow with a high level of performance in identifying and sizing branches of apple trees in a dynamic orchard environment and integrating farm management practices into automated decision-making for optimizing crop-load in apple orchards.
Why it matches plant phenotyping methodsRGB-D画像とYOLOv8・PCAを用いてリンゴ枝径および枝ごとの作物負荷を推定する手法を開発・評価しており、植物形質の取得が中心です。
abstractIn this study, we propose a machine vision system for estimating one of the canopy parameters in apple orchards: branch diameter.
Maize leaf area offers valuable insights into physiological processes, playing a critical role in breeding and guiding agricultural practices. The Azure Kinect DK possesses the real-time capability to capture and analyze the spatial structural features of crops. However, its further application in maize leaf area measurement is constrained by RGB–depth misalignment and limited sensitivity to detailed organ-level features. This study proposed a novel approach to address and optimize the limitations of the Azure Kinect DK through the multimodal coupling of RGB-D data for enhanced organ-level crop phenotyping. To correct RGB–depth misalignment, a unified recalibration method was developed to ensure accurate alignment between RGB and depth data. Furthermore, a semantic information-guided depth inpainting method was proposed, designed to repair void and flying pixels commonly observed in Azure Kinect DK outputs. The semantic information was extracted using a joint YOLOv11-SAM2 model, which utilizes supervised object recognition prompts and advanced visual large models to achieve precise RGB image semantic parsing with minimal manual input. An efficient pixel filter-based depth inpainting algorithm was then designed to inpaint void and flying pixels and restore consistent, high-confidence depth values within semantic regions. A validation of this approach through leaf area measurements in practical maize field applications—challenged by a limited workspace, constrained viewpoints, and environmental variability—demonstrated near-laboratory precision, achieving an MAPE of 6.549%, RMSE of 4.114 cm2, MAE of 2.980 cm2, and R2 of 0.976 across 60 maize leaf samples. By focusing processing efforts on the image level rather than directly on 3D point clouds, this approach markedly enhanced both efficiency and accuracy with the sufficient utilization of the Azure Kinect DK, making it a promising solution for high-throughput 3D crop phenotyping.
Why it matches plant phenotyping methodsAzure Kinect RGB-D再較正、深度補間、意味解析を開発し、トウモロコシ葉面積測定で検証した、植物表現型取得法が中心の研究。
abstractThis study proposed a novel approach to address and optimize the limitations of the Azure Kinect DK through the multimodal coupling of RGB-D data for enhanced organ-level crop phenotyping.
Field / plotLiDAR / point cloudRGB-D / ToFStem / branchMorphology / geometry measurementObject detectionSegmentationArchitecture / morphology / geometry
Abstract Green asparagus is a labor-intensive vegetable crop for harvesting. Rising labor costs and seasonal labor shortages are threatening the sustainability and profitability of the asparagus industry in the United States (U.S.), making harvesting automation more urgently needed than ever. A multitude of challenges, however, exist in developing practically variable automated harvesting technology for green asparagus. This study describes a novel effort aimed at the development and evaluation of a 3D vision-based asparagus perception system for selective harvesting of U.S. green asparagus. Three different types of 3D cameras were evaluated indoors for asparagus detection and length estimation. A time-of-flight technology-based camera yielded the best accuracy and was deployed on a mobile vision system for imaging asparagus in diverse field conditions. A new annotated dataset alongside ground-truth length measurements for 1008 spears was created for asparagus perception algorithm development and evaluation. Among three small-scale YOLO (v8, v9, and v10) models, YOLOv8s achieved the best F1 score of 0.849. The YOLOv8s-based pipeline point cloud processing algorithms including segmentation, clustering, and outlier removal, achieved a percentage accuracy of 89.4% (with the corresponding error of about 2.3 cm) in spear length estimation, and the averaged base point localization error of 1.4 cm. The entire algorithm pipeline for spear detection and localization could be run at about 3.3 frames per second on an onboard computer. The harvest eligibility analysis of detected spears showed a detection rate of 95.8% of harvestable spears when only the length criterion (greater than 8 inches or 20.3 cm) was applied and a detection rate of 84.5% when both the length and base point localization (error less than 2 cm) criteria were considered. The 3D vision-based perception system in this study is promising for reliable and real-time asparagus perception in production fields.
Why it matches plant phenotyping methods3Dビジョンによるアスパラガスの検出・長さ推定システムを開発・評価し、植物器官形質の測定精度とデータセットを提示しているため、収穫ロボット用の単なる位置検出を超えた中心的な表現型計測研究である。
abstractThis study describes a novel effort aimed at the development and evaluation of a 3D vision-based asparagus perception system for selective harvesting of U.S. green asparagus.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 15 Sept 2026
Field / plotRGB-D / ToFStem / branchObject detection2D/3D reconstructionSegmentation
Green asparagus has the characteristic of growing in clusters, making it inevitable for harvest targets to overlap with weeds and immature asparagus in the field. Extracting stem details in complex spatial positions information presents a significant challenge in identifying suitable harvest targets and high-precision cutting-points. This paper explored YS3AM (Yolo-SAM-3D-Adaptive-Modeling) method for green asparagus detection and 3D adaptive-section modeling using a depth camera, which could furnish harvesting path planning for the selective harvesting robots. Firstly, the model was developed and deployed to extract bounding boxes for individual asparagus stems within clusters. Secondly, the green asparagus stems within these bounding boxes were segment and generate binary mask images. Thirdly, high-quality depth images were obtained using pixel block completion. Finally, based on the cylinder, an adaptive-section 3D reconstruction method fusion with mask and depth was proposed, with a novel evaluation method applied to assess modeling accuracy. The experimental detection results of 1,095 test images demonstrated that the Precision was 98.75%, the Recall was 95.46%, the F1 score was 0.97, and the mAP was 97.16%. The modeling accuracy of 103 asparagus stems under sunny (54) and cloudy (49) conditions was estimated. The average RMSEs of length and bottom depth were 0.74 and 1.105. The detection and modeling for each stem approximately demanded 22 ms. The results of this paper indicated that the 3D model effectively represented the spatial distribution of green asparagus, and further accurately identification of suitable harvest targets and stem cutting-points. This model provided essential spatial pathways for end-effector path planning, thereby fulfilling the operational requirements for efficient green asparagus harvesting robot.
Why it matches plant phenotyping methodsアスパラガス茎の検出に加え、深度画像とマスクから茎の3D形状を再構成し、長さ・底部深度を定量化する手法が中心である。収穫対象の単なる位置検出を超えた器官形質の推定と精度評価を含むため採用。
abstractThis paper explored YS3AM (Yolo-SAM-3D-Adaptive-Modeling) method for green asparagus detection and 3D adaptive-section modeling using a depth camera
High-throughput phenotyping (HTP) has become a powerful tool for gaining insights into the genetic and environmental factors that affect cotton (Gossypium spp.) growth and yield. With the recent advances in the field of computer vision, namely the integration of deep learning algorithms, the accuracy and efficiency of HTP systems have improved dramatically, enabling them to automatically quantify such fundamental phenotypic traits as fruit identification and enumeration. However, there is currently no HTP system available for counting all the reproductive phases of cotton crop that can be deployed in agronomic field conditions throughout the growing season. This study presents CottonSense, an advanced HTP system that overcomes the challenges of deployment across multiple growth periods by effectively segmenting and enumerating cotton fruits at four stages of growth, including square, flower, closed boll, and open boll. Consequently, CottonSense enhances agronomic management through increased opportunities for data collection and analysis. Using RGB-D cameras, it captures and processes both two and three-dimensional data, facilitating a wider range of phenotypic trait extractions such as crop biomass and plant architecture. To segment the cotton fruits, a Mask-RCNN model is trained and optimized for faster inference using TensorRT. The model yields an average AP score of 79% in segmentation across the four fruit categories. Moreover, the model's accuracy in estimating total fruit count per image is validated by a strong agreement with the counts given by ten domain experts, as reflected by an R² value of 0.94. Furthermore, to accurately count the segmented fruits over large populations of plants, an enumeration algorithm based on a tracking strategy is developed that achieves an R² value of 0.93 when compared to hand-counted fruits in the field. The proposed HTP system, which is implemented entirely on an edge computing device, is cost-effective and power-efficient, making it an effective tool for high-yield cotton breeding and crop improvement. The code for CottonSense is publicly available at https://github.com/FeriBolour/CottonSense.
Why it matches plant phenotyping methods綿花の果実を画像から分割・計数する高スループット表現型計測システムを開発し、専門家および手作業計数で検証しているため、方法が研究の中心である。
abstractThis study presents CottonSense, an advanced HTP system that overcomes the challenges of deployment across multiple growth periods by effectively segmenting and enumerating cotton fruits at four stages of growth, including square, flower, closed boll, and open boll.
The compact, high-throughput phenotyping platform, characterized by its portability and small size, is well-suited for crop phenotyping across diverse environments. However, integrating multi-source sensors to achieve synchronized data acquisition and analysis poses significant challenges due to constraints in load capacity and available space. To address these issues, we developed a robotic platform specifically designed for phenotyping greenhouse strawberries. This system integrates an RGB-D camera, a multispectral camera, a thermal camera, and a LiDAR sensor, enabling the unified analysis of data from these sources. The platform accurately extracted key phenotypic parameters, including canopy width (R² = 0.9864, RMSE = 0.0185 m) and average temperature (R2 = 0.8056, RMSE = 0.1732 °C), with errors maintained below 5%. Furthermore, it effectively distinguished between different strawberry varieties, achieving an Adjusted Rand Index of 0.94, underscoring the value of detailed phenotyping in variety differentiation. Compared to conventional UGV-LiDAR systems, the proposed platform is more cost-effective, efficient, and scalable, with enhanced data consistency, making it a promising solution for agricultural applications.
Why it matches plant phenotyping methods複数センサーを統合したロボット植物フェノタイピング基盤を開発し、キャノピー幅と平均温度の抽出精度を検証しているため、フェノタイピング手法が中心である。
abstractwe developed a robotic platform specifically designed for phenotyping greenhouse strawberries.
Potato yield is an important metric for farmers to further optimize their cultivation practices. Potato yield can be estimated on a harvester using an RGB-D camera that can estimate the three-dimensional (3D) volume of individual potato tubers. A challenge, however, is that the 3D shape derived from RGB-D images is only partially completed, underestimating the actual volume. To address this issue, we developed a 3D shape completion network, called CoRe++, which can complete the 3D shape from RGB-D images. CoRe++ is a deep learning network that consists of a convolutional encoder and a decoder. The encoder compresses RGB-D images into latent vectors that are used by the decoder to complete the 3D shape using the deep signed distance field network (DeepSDF). To evaluate our CoRe++ network, we collected partial and complete 3D point clouds of 339 potato tubers on an operational harvester in Japan. On the 1425 RGB-D images in the test set (representing 51 unique potato tubers), our network achieved a completion accuracy of 2.8 mm on average. For volumetric estimation, the root mean squared error (RMSE) was 22.6 ml, and this was better than the RMSE of the linear regression (31.1 ml) and the base model (36.9 ml). We found that the RMSE can be further reduced to 18.2 ml when performing the 3D shape completion in the center of the RGB-D image. With an average 3D shape completion time of 10 ms per tuber, we can conclude that CoRe++ is both fast and accurate enough to be implemented on an operational harvester for high-throughput potato yield estimation. CoRe++’s high-throughput and accurate processing allows it to be applied to other tuber, fruit and vegetable crops, thereby enabling versatile, accurate and real-time yield monitoring in precision agriculture. Our code, network weights and dataset are publicly available at https://github.com/UTokyo-FieldPhenomics-Lab/corepp.git.
Why it matches plant phenotyping methodsRGB-D画像からジャガイモ塊茎の3D形状と体積を推定する手法を開発・検証し、収穫量フェノタイピングに適用した中心的な方法論研究。
abstractwe developed a 3D shape completion network, called CoRe++, which can complete the 3D shape from RGB-D images.
We address the problem of agricultural image segmentation by introducing a novel loss formulation called Adaptive Select Loss (ASL), inspired by the Top-k loss strategy. While Top-k loss was originally designed for classification tasks, ASL is specifically tailored for semantic segmentation. It exploits the hierarchical structure of loss computation specific in semantic segmentation–aggregated first at the pixel level, then at the image level–while accounting for the imbalance between precise but scarce image-level annotations and noisy yet abundant pixel-level labels. ASL selectively aggregates loss from a “Few” (Top-k) of the most informative image-level instances and from “Almost all” (remove few) pixel-level data, thereby balancing robustness and sensitivity to noise. To ensure stability during training, we introduce a derivative smoothing mechanism that addresses the convergence issues introduced by the hard selection threshold, particularly when training with small number of images for loss aggregation. Empirically, the proposed approach improves boundary localization and segmentation quality in the presence of annotation noise. We evaluate ASL on three challenging semantic segmentation tasks–two agricultural and one mixed–using a visual transformer backbone, including hyperspectral data. ASL achieves consistent performance improvements, with gains of approximately 2.5% on hyperspectral and satellite imagery, and up to 6% on RGB-D data plant segmentation problem.
Why it matches plant phenotyping methods農業画像の植物セグメンテーション性能を改善する損失関数を開発し、植物セグメンテーションを含む複数課題で検証しており、画像ベースの植物状態抽出手法が中心である。
abstractWe address the problem of agricultural image segmentation by introducing a novel loss formulation called Adaptive Select Loss (ASL)
Selective harvesting robots for broccoli face significant challenges in field operations, where occlusions by leaves and stems, varying maturity stages and lighting interferences greatly affect performance. Addressing the need for a robust network capable of maturity recognition and localisation under various occlusion conditions for spherical crops, OccluInst-a single-stage instance segmentation network based on RGB-D and CNN-Transformer architecture was proposed. The solution is to make full use of visible information and crop characteristics. This model builds a dual-branch cross-modal calibration framework to generate instance-aware kernels and segmentation mask features. The proposed Attention Weight Interactive Fusion Module (AWIF) enhances the fusion efficiency of multi-scale RGB and depth features in complex scenarios, while the designed Adaptive Fusion Ratio Module (AFR) filters out noisy depth data and extracts valuable information to achieve feature alignment. Additionally, the developed Material Awareness Module (MA) highlights critical areas, improving feature extraction for irregular, multi-scale targets. The improved circular boundary anchor box accurately localises broccoli under various levels of occlusion. Ablation studies confirm the effectiveness of each module. OccluInst can swiftly and accurately identify the maturity categories and coordinates of broccoli under different occlusion levels. It achieves a mAP₅₀ of 86.2% and mAR of 83.5%, with an average centre point deviation of 3.68 pixels on images with a resolution of 848×480, and a detection speed of 51.4 frames per second, providing a robust visual foundation for selective harvesting robots.
Why it matches plant phenotyping methodsRGB-D画像からブロッコリーの成熟カテゴリーという植物状態を推定するセグメンテーション手法を開発し、遮蔽条件下で性能評価している。単なる収穫対象の位置検出にとどまらず、成熟度推定が中心的である。
abstractOccluInst can swiftly and accurately identify the maturity categories and coordinates of broccoli under different occlusion levels.
Fruit size is crucial for growers as it influences consumer willingness to buy and the price of the fruit. Fruit size and growth along the seasons are two parameters that can lead to more precise orchard management favoring production sustainability. In this study, a Python-based computer vision system (CVS) for sizing apples directly on the tree was developed to ease fruit sizing tasks. The system is made of a consumer-grade depth camera and was tested at two distances among 17 timings throughout the season, in a Fuji apple orchard. The CVS exploited a specifically trained YOLOv5 detection algorithm, a circle detection algorithm, and a trigonometric approach based on depth information to size the fruits. Comparisons with standard-trained YOLOv5 models and with spherical objects were carried out. The algorithm showed good fruit detection and circle detection performance, with a sizing rate of 92%. Good correlations (r > 0.8) between estimated and actual fruit size were found. The sizing performance showed an overall mean error (mE) and RMSE of + 5.7 mm (9%) and 10 mm (15%). The best results of mE were always found at 1.0 m, compared to 1.5 m. Key factors for the presented methodology were: the fruit detectors customization; the HoughCircle parameters adaptability to object size, camera distance, and color; and the issue of field natural illumination. The study also highlighted the uncertainty of human operators in the reference data collection (5–6%) and the effect of random subsampling on the statistical analysis of fruit size estimation. Despite the high error values, the CVS shows potential for fruit sizing at the orchard scale. Future research will focus on improving and testing the CVS on a large scale, as well as investigating other image analysis methods and the ability to estimate fruit growth.
Why it matches plant phenotyping methods果実サイズという植物器官形質を、深度カメラ・物体検出・円検出・三角測量で推定するコンピュータビジョン手法を開発・検証しており、フェノタイピング手法が中心である。
abstracta Python-based computer vision system (CVS) for sizing apples directly on the tree was developed
Reproduction assets foundThe paper's RGB-D apple dataset (RGB/depth frames, annotations, and reference caliper measurements) is explicitly released as open source on GitHub with a Zenodo DOI. The YOLOv5 base model is a generic third-party library, not a paper-specific asset; no author analysis code repository is stated.Dataset · publicThe obtained dataset is open source and available at
https://github.com/ECOPOM/OpenAcces_RGBD_apple_dataset (Bortolotti et al., 2024).Open asset ↗ECOPOM/OpenAcces_RGBD_apple_datasetpdf-page:3 lines:1-52Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Published1 Dec 2024Computers and Electronics in Agriculture.
The three-dimensional reconstruction of fruit trees plays a crucial role in assessing their growth status, analyzing agronomic traits, and categorizing their organs. This is vital for implementing intelligent orchard management. This study aims to develop a cost-effective and efficient method for the three-dimensional reconstruction and skeleton extraction of fruit trees. The proposed method leverages the 3D geometric structure captured by Time-of-Flight (TOF) sensors and addresses common issues such as occlusion and perspective ambiguity. Firstly, the TOF sensor and its supporting components are used to build an acquisition platform to collect the full range point cloud of fruit trees in the key growth period. The noise information is filtered through the point cloud preprocessing operation to obtain the complete target point cloud and extract its structural invariant features. The IWOA-RANSAC-NDT algorithm is introduced for 3D model registration. Secondly, the Delaunay triangulation algorithm and Dijkstra shortest path algorithm are used to calculate the Minimum Spanning Tree. Branch segmentation is expedited using the Kd-tree data structure. The Levenberg Marquardt algorithm and the cylindrical fitting method are used to obtain the full fruit tree skeleton model. Finally, taking walnut tree as the experimental object, a high-precision fruit tree point cloud model is constructed, and the actual verification is carried out based on the measured data. Findings indicate that the proposed methodology can accurately construct both 3D point cloud and skeleton models of fruit trees with accuracy deviations from the measured data remaining within 7 %. The proposed method offers valuable data and technical support for the future development of highly autonomous, practical, and user-oriented fruit tree pruning systems.
Why it matches plant phenotyping methods果樹の3D形態・骨格を取得および抽出するTOFセンサベースの手法と取得プラットフォームを開発し、実測データで検証しているため、植物フェノタイピング手法が中心である。
abstractThis study aims to develop a cost-effective and efficient method for the three-dimensional reconstruction and skeleton extraction of fruit trees.
Intelligent pruning is an effective way to improve the efficiency of fruit tree pruning and reduce production costs, where the positioning of fruit tree pruning points is the key. In this study, a pruning point localisation method based on deep learning and red-green-blue-depth (RGB-D) is proposed for dormant tall spindle apple trees, which can quickly and accurately identify the pruning points on the primary branches. Firstly, red-green-blue (RGB) images and depth images of apple trees were acquired by using the Realsense D435i depth camera, and the SOLOv2 instance segmentation model was used to segment the trunks, branches, and supports in RGB images. Secondly, the manual pruning rules were adapted to improve the pruning methods; then using OpenCV image processing method, the junction points of branches and trunk and potential pruning points were gained, and the world coordinates of the points were obtained according to the coordinate transformation to calculate the length of branch diameter and spacing. Finally, the pruning points were determined according to the pruning rules. The results show that SOLOv2 has better segmentation effect compared with Mask R–CNN and Cascade Mask R–CNN. The mean absolute error between the estimated and manually measured values of branch diameter and spacing are 1.10 mm and 16.06 mm, and the recognition accuracy of pruning points is 87.2%, with a recognition time of about 3.8 s for each image. It is shown that the method can measure branch diameter and spacing and quickly locate pruning points with high reliability and accuracy, and the study provides a basis for the development of apple tree pruning robots.
Why it matches plant phenotyping methodsRGB-D画像と画像処理・深層学習による枝径・枝間隔の推定および剪定点抽出が研究の中心であり、植物形態形質の取得手法を開発・検証しているため。
abstracta pruning point localisation method based on deep learning and red-green-blue-depth (RGB-D) is proposed
Potato yield is an important metric for farmers to further optimize their cultivation practices. Potato yield can be estimated on a harvester using an RGB-D camera that can estimate the three-dimensional (3D) volume of individual potato tubers. A challenge, however, is that the 3D shape derived from RGB-D images is only partially completed, underestimating the actual volume. To address this issue, we developed a 3D shape completion network, called CoRe++, which can complete the 3D shape from RGB-D images. CoRe++ is a deep learning network that consists of a convolutional encoder and a decoder. The encoder compresses RGB-D images into latent vectors that are used by the decoder to complete the 3D shape using the deep signed distance field network (DeepSDF). To evaluate our CoRe++ network, we collected partial and complete 3D point clouds of 339 potato tubers on an operational harvester in Japan. On the 1425 RGB-D images in the test set (representing 51 unique potato tubers), our network achieved a completion accuracy of 2.8 mm on average. For volumetric estimation, the root mean squared error (RMSE) was 22.6 ml, and this was better than the RMSE of the linear regression (31.1 ml) and the base model (36.9 ml). We found that the RMSE can be further reduced to 18.2 ml when performing the 3D shape completion in the center of the RGB-D image. With an average 3D shape completion time of 10 ms per tuber, we can conclude that CoRe++ is both fast and accurate enough to be implemented on an operational harvester for high-throughput potato yield estimation. CoRe++’s high-throughput and accurate processing allows it to be applied to other tuber, fruit and vegetable crops, thereby enabling versatile, accurate and real-time yield monitoring in precision agriculture. Our code, network weights and dataset are publicly available at https://github.com/UTokyo-FieldPhenomics-Lab/corepp.git . • CoRe++ is a high-throughput 3D shape completion network for RGB-D images. • CoRe++ uses a convolutional encoder and DeepSDF decoder to complete the 3D shape. • Testing on 1425 RGB-D images, CoRe++ achieved a completion accuracy of 2.8 mm. • The average 3D shape completion time was 10 ms per potato tuber.
Why it matches plant phenotyping methodsRGB-D画像からジャガイモ塊茎の3D形状と体積を推定する手法を開発・評価しており、植物形質取得が研究の中心である。
abstractwe developed a 3D shape completion network, called CoRe++, which can complete the 3D shape from RGB-D images.
Realizing the integration of intelligent fruit picking and grading for apple harvesting robots is an inevitable requirement for the future development of smart agriculture and precision agriculture. Therefore, an apple maximum diameter estimation model based on RGB-D camera fusion depth information was proposed in the study. Firstly, the maximum diameter parameters of Red Fuji apples were collected, and the results were statistically analyzed. Then, based on the Intel RealSense D435 RGB-D depth camera and LabelImg software, the depth information of apples and the two-dimensional size information of fruit images were obtained. Furthermore, the relationship between fruit depth information, two-dimensional size information of fruit images, and the maximum diameter of apples was explored. Based on Origin software, multiple regression analysis and nonlinear surface fitting were used to analyze the correlation between fruit depth, diagonal length of fruit bounding rectangle, and maximum diameter. A model for estimating the maximum diameter of apples was constructed. Finally, the constructed maximum diameter estimation model was experimentally validated and evaluated for imitation apples in the laboratory and fruits on the Red Fuji fruit trees in modern apple orchards. The experimental results showed that the average maximum relative error of the constructed model in the laboratory imitation apple validation set was ±4.1%, the correlation coefficient (R2) of the estimated model was 0.98613, and the root mean square error (RMSE) was 3.21 mm. The average maximum diameter estimation relative error on the modern orchard Red Fuji apple validation set was ±3.77%, the correlation coefficient (R2) of the estimation model was 0.84, and the root mean square error (RMSE) was 3.95 mm. The proposed model can provide theoretical basis and technical support for the selective apple-picking operation of intelligent robots based on apple size grading.
Why it matches plant phenotyping methodsRGB-D画像と深度情報からリンゴ果実の最大径を推定する手法を開発し、実験室および果樹園で検証しており、収穫ロボット用途を超えて果実形質の取得が中心である。
abstractan apple maximum diameter estimation model based on RGB-D camera fusion depth information was proposed in the study
Automated phenotyping is the task of automatically measuring plant attributes to help farmers and breeders in developing and growing strong robust plants. An automated tool for early illness detection can accelerate the process of identifying plant resistance and quickly pinpoint problematic breeding. Many such phenotyping tasks can be achieved by analyzing images from simple, low cost, RGB-D sensors. In this paper we focused on a particular case study — identifying the resistance level of tomato hybrids to the tomato yellow leaf curl virus (TYLCV) in production greenhouses. This is a difficult task, as separating between resistance levels based on images is difficult even for expert breeders. We collected a large dataset of images from an experiment containing many tomato hybrids with varying resistance levels. We used the depth information to identify the topmost part of the tomato plant. We then used deep learning models to classify the various resistance levels. For identifying plants with visual symptoms, our methods achieved an accuracy of 0.928, a precision of 0.934, and a recall of 0.95. In the multi-class case we achieved an accuracy of 0.76 in identifying the correct level, and an error of 0.278. Our methods are not particularly tailored for the specific task, and can be extended to other tasks that identify various plant diseases with visual symptoms such as ToBRFV, mildew, ToMV and others.
Why it matches plant phenotyping methodsRGB-D画像と深度情報、深層学習を用いてトマトのウイルス症状と抵抗性レベルを推定する手法を開発・評価しており、植物表現型取得が中心である。
abstractAutomated phenotyping is the task of automatically measuring plant attributes
In this study, a robust method for 3D pose estimation of immature green apples (fruitlets) in commercial orchards was developed, utilizing the YOLO11 object pose detection model alongside Vision Transformers (ViT) for depth estimation. For object detection and pose estimation, performance comparisons of YOLO11 (YOLO11n, YOLO11s, YOLO11m, YOLO11l and YOLO11x) and YOLOv8 (YOLOv8n, YOLOv8s, YOLOv8m, YOLOv8l and YOLOv8x) were made under identical hyperparameter settings among the all configurations. Likewise, for RGB to RGB-D mapping, Dense Prediction Transformer (DPT) and Depth Anything V2 were investigated. It was observed that YOLO11n surpassed all configurations of YOLO11 and YOLOv8 in terms of box precision and pose precision, achieving scores of 0.91 and 0.915, respectively. Conversely, YOLOv8n exhibited the highest box and pose recall scores of 0.905 and 0.925, respectively. Regarding the mean average precision at 50% intersection over union (mAP@50), YOLO11s led all configurations with a box mAP@50 score of 0.94, while YOLOv8n achieved the highest pose mAP@50 score of 0.96. In terms of image processing speed, YOLO11n outperformed all configurations with an impressive inference speed of 2.7 ms, significantly faster than the quickest YOLOv8 configuration, YOLOv8n, which processed images in 7.8 ms. This demonstrates a substantial improvement in inference speed over previous iterations, particularly evident when comparing YOLO11n and YOLOv8n. Subsequent integration of ViTs for the green fruit's pose depth estimation revealed that Depth Anything V2 outperformed Dense Prediction Transformer in 3D pose length validation, achieving the lowest Root Mean Square Error (RMSE) of 1.52 and Mean Absolute Error (MAE) of 1.28, demonstrating exceptional precision in estimating immature green fruit lengths. Following this, the DPT showed notable accuracy improvements with a RMSE of 3.29 and an MAE of 2.62. In contrast, measurements derived from Intel RealSense point clouds exhibited the highest discrepancies from the ground truth, with a RMSE of 9.98 and an MAE of 7.74. These findings emphasize the effectiveness of YOLO11 in detecting and estimating the pose of immature green fruits, illustrating how Vision Transformers like Depth Anything V2 adeptly convert RGB images into RGB-D data, thus enhancing the precision and computational requirement of 3D pose estimations for future robotic thinning applications in commercial orchards.
Why it matches plant phenotyping methods未熟果実の検出にとどまらず、3D姿勢と果実長を推定し、複数モデルを比較・検証する画像ベースの植物器官形質計測法が中心である。
abstracta robust method for 3D pose estimation of immature green apples (fruitlets) in commercial orchards was developed
Accurately detecting the maturity and 3D position of flowering Chinese cabbage ( Brassica rapa var. chinensis) in natural environments is vital for autonomous robot harvesting in unstructured farms. The challenge lies in dense planting, small flower buds, similar colors and occlusions. This study proposes a YOLOv8-Improved network integrated with the ByteTrack tracking algorithm to achieve multi-object detection and 3D positioning of flowering Chinese cabbage plants in fields. In this study, C2F-MLCA is created by adding a lightweight Mixed Local Channel Attention (MLCA) with spatial awareness capability to the C2F module of YOLOv8, which improves the extraction of spatial feature information in the backbone network. In addition, a P2 detection layer is added to the neck network, and BiFPN is used instead of PAN to enhance multi-scale feature fusion and small target detection. Wise-IoU in combination with Inner-IoU is adopted as a new loss function to optimize the network for different quality samples and different size bounding boxes. Lastly, ByteTrack is integrated for video tracking, and RGB-D camera depth data are used to estimate cabbage positions. The experimental results show that YOLOv8-Improve achieves a precision ( P ) of 86.5% and a recall ( R ) of 86.0% in detecting the maturity of flowering Chinese cabbage. Among them, mAP50 and mAP75 reach 91.8% and 61.6%, respectively, representing an improvement of 2.9% and 4.7% over the original network. Additionally, the number of parameters is reduced by 25.43%. In summary, the improved YOLOv8 algorithm demonstrates high robustness and real-time detection performance, thereby providing strong technical support for automated harvesting management.
Why it matches plant phenotyping methods開花中国白菜の成熟度という植物状態を画像から検出・推定する改良YOLOv8と3D位置推定手法が研究の中心であり、収穫対象の単なる位置検出を超える。
abstractThis study proposes a YOLOv8-Improved network integrated with the ByteTrack tracking algorithm to achieve multi-object detection and 3D positioning of flowering Chinese cabbage plants in fields.
Alfalfa is a high-yielding forage crop that is widely grown in the United States for grazing, hay and silage making. A proper maintenance of these grasslands is necessary to ensure optimum productivity and profits. The pre-harvest estimation of biomass yield helps in quantifying the profits and optimizing the forage allocation in advance. Most traditional methods of forage estimation are relatively laborious and time-consuming. Recent developments in contact and remote sensing technologies opened numerous paths for performing aboveground biomass estimation tasks with flexibility and easiness. This study focused on the development of crop height measurement systems for estimating the aboveground biomass yield of Alfalfa (Medicago sativa L.). Five different systems for measuring crop height were evaluated on their ability to estimate aboveground wet and dry biomass. The crop height measurement systems used in this study were Structure-from-Motion, Ultrasound Sensor and Ski, Inertial Measurement Unit and Ski, Inertial Measurement Units and Roller, and a Depth Camera. The results indicated that the system using the Inertial Measurement Unit sensor and ski (IMU-Ski) performed the best among ground-based methods (R2= 0.79; SeY= 3166 kg-wet/ha). The Structure-from-Motion (SfM) method using UAV also provided satisfactory results for biomass predictions (R2= 0.74; SeY= 2543 kg-wet/ha). The models based on IMU-Ski and UAV-based SfM methods were facilitated with vegetation coverage as an additional independent variable to evaluate their effect on biomass predictions. The results indicated that the vegetation coverage did not improve the predictions in any of these systems. Thus, the models based on only the crop height (IMU-Ski and UAV-SfM) were the recommended approaches for Alfalfa biomass estimations. The addition of data points for wide ranges of crop height and vegetation coverage is recommended for future studies to improve the results and ensure the adaptability of these systems in varying environmental conditions.
Why it matches plant phenotyping methodsアルファルファの草丈から地上部バイオマスを推定する複数のフィールドロボティクス・センシング手法を開発・比較評価しており、植物形質の取得と推定が研究の中心である。
abstractThis study focused on the development of crop height measurement systems for estimating the aboveground biomass yield of Alfalfa (Medicago sativa L.).
Monitoring the number of cotton flowers can provide important information for breeders to assess the flowering time and the productivity of genotypes because flowering marks the transition from vegetative growth to reproductive growth and impacts the final yield. Traditional manual counting methods are time-consuming and impractical for large-scale fields. To count cotton flowers efficiently and accurately, a multi-view multi-object tracking approach was proposed by using both RGB and depth images collected by three RGB-D cameras fixed on a ground robotic platform. The tracking-by-detection algorithm was employed to track flowers from three views simultaneously and remove duplicated counting from single views. Specifically, an object detection model (YOLOv8) was trained to detect flowers in RGB images and a deep learning-based optical flow model Recurrent All-pairs Field Transforms (RAFT) was used to estimate motion between two adjacent frames. The intersection over union and distance costs were employed to associate flowers in the tracking algorithm. Additionally, tracked flowers were segmented in RGB images and the depth of each flower was obtained from the corresponding depth image. Those flowers tracked with known depth from two side views were then projected onto the middle image coordinate using camera calibration parameters. Finally, a constrained hierarchy clustering algorithm clustered all flowers in the middle image coordinate to remove duplicated counting from three views. The results showed that the mean average precision of trained YOLOv8x was 96.4%. The counting results of the developed method were highly correlated with those counted manually with a coefficient of determination of 0.92. Besides, the mean absolute percentage error of all 25 testing videos was 6.22%. The predicted cumulative flower number of Pima cotton flowers is higher than that of Acala Maxxa, which is consistent with what breeders have observed. Furthermore, the developed method can also obtain the flower number distributions of different genotypes without laborious manual counting in the field. Overall, the three-view approach provides an efficient and effective approach to count cotton flowers from multiple views. By collecting the video data continuously, this method is beneficial for breeders to dissect genetic mechanisms of flowering time with unprecedented spatial and temporal resolution, also providing a means to discern genetic differences in fecundity, the number of flowers that result in harvestable bolls. The code and datasets used in this paper can be accessed on GitHub: https://github.com/UGA-BSAIL/Multi-view_flower_counting.
Why it matches plant phenotyping methodsRGB-Dカメラ、ロボットプラットフォーム、物体追跡、深度投影、重複除去を統合してワタ花数を自動計測する手法を開発・検証しており、植物表現型取得が中心である。
abstractTo count cotton flowers efficiently and accurately, a multi-view multi-object tracking approach was proposed by using both RGB and depth images collected by three RGB-D cameras fixed on a ground robotic platform.
Drought is a main abiotic stress facing agriculture and forestry production and its impacts are exacerbated by climate change. Accurately and effectively monitoring the drought stress levels of crop and tree species is crucial for their efficient management and the selection of drought-resistant varieties. This study used four poplar seedlings with differing drought tolerance and conducted five drought stress level tests. A self-propelled phenotyping platform was constructed and equipped with an Intel RealSense D435i RGB-D (Red-Green- Blue-Depth) camera and a RedEdge-MX multispectral camera. The side-view RGB and depth images and five-channel top-view multispectral images of poplar seedlings were collected by this platform; and the color, texture, depth, and spectral features were extracted through image processing. In addition, plant height, ground diameter, leafstalk angle, chlorophyll content and water content were collected from the poplar seedlings. Using long short-term memory (LSTM), a multi-output classification was performed on the varieties and drought stress levels of the four poplar seedlings based on the 50 parameters obtained through manual and image processing. By combining ResNet18 and the improved ResNet18 embedded in the convolutional block attention module (CBAM) with the LSTM model, the resulting ResNet18-LSTM and ResNet18-CBAM-LSTM models were used to perform multi-output grading on the varieties and drought stress levels. As for varieties, the classification accuracy of LSTM, ResNet18-LSTM, and ResNet18-CBAM-LSTM models were 96.56 %, 98.12 %, and 99.69 %, respectively. As for drought stress levels, the classification accuracy of the three models were 83.44 %, 85.62 %, and 90.94 %, respectively. The ResNet18-CBAM-LSTM model performed the best in the classification of the two parameters. This study comprehensively and continuously monitored the dynamic response of multiple varieties of poplar seedlings under different drought conditions, and a new perspective for the classification of drought stress levels and the breeding of better varieties are provided.
Why it matches plant phenotyping methods自走式フェノタイピングプラットフォーム、RGB-D・マルチスペクトル画像、画像処理による形質抽出、深層学習分類が研究の中心であり、ポプラの品種と干ばつストレス状態を評価する実質的なフェノタイピング手法研究である。
abstractA self-propelled phenotyping platform was constructed and equipped with an Intel RealSense D435i RGB-D (Red-Green- Blue-Depth) camera and a RedEdge-MX multispectral camera.
In modern agriculture, precise monitoring of plants and fruits is crucial for tasks such as high-throughput phenotyping and automated harvesting. This paper addresses the challenge of reconstructing accurate 3D shapes of fruits from partial views, which is common in agricultural settings. We introduce CF-PRNet, a coarse-to-fine prototype refining network, leverages high-resolution 3D data during the training phase but requires only a single RGB-D image for real-time inference. Our approach begins by extracting the incomplete point cloud data that constructed from a partial view of a fruit with a series of convolutional blocks. The extracted features inform the generation of scaling vectors that refine two sequentially constructed 3D mesh prototypes - one coarse and one fine-grained. This progressive refinement facilitates the detailed completion of the final point clouds, achieving detailed and accurate reconstructions. CF-PRNet demonstrates excellent performance metrics with a Chamfer Distance of 3.78, an F1 Score of 66.76%, a Precision of 56.56%, and a Recall of 85.31%, and win the first place in the Shape Completion and Reconstruction of Sweet Peppers Challenge.
Why it matches plant phenotyping methods果実の部分RGB-D画像から3D形状を再構成する手法を開発・評価しており、植物器官の形態形質取得が研究の中心である。
abstractThis paper addresses the challenge of reconstructing accurate 3D shapes of fruits from partial views
Reproduction assets foundThe paper's authors publicly release their CF-PRNet source code for sweet pepper point cloud completion. The sweet pepper benchmark dataset is cited prior work (ref [2]), not a paper-specific asset, and the challenge website is a generic event page.Code · publicOur source code is
available at https://github.com/uqzhichen/CF-PRNet/.Open asset ↗uqzhichen/CF-PRNetpdf-page:1 lines:1-50Plant phenotyping relevance match · UnverifiedOpenAlex · checked 15 Sept 2026
Trunk diameter is related to the overall health and level of carbon sequestration in a tree. Trunk diameter measurement, therefore, is a key task in both forest plot and urban settings. Unlike the traditional approach of manual measurement with a measuring tape or calipers, several recent approaches rely on sophisticated technologies such as LiDAR and time-of-flight cameras that provide fine-grain depth maps, which are used for depth-assisted image segmentation in downstream processing. These technologies are supported only on specialized devices or high-end smartphones. We present a mobile application that uses coarse-grain depth maps derived from an optical sensor, and so can be run on most common Android devices. Moreover, we use a state-of-the-art deep neural network to estimate trunk diameter from an image and its corresponding coarse depth map (RGB-D). We tested our app using a data set collected from four countries and under challenging conditions including occlusion, leaning trees, and irregular shapes and found that our algorithm has a MAE of 1.66 cm and an RMSE of 2.46 cm, which is comparable to accuracy from fine-grain depth maps. Moreover, diameter measurement using our app is >5 times faster than traditional manual surveying.
Why it matches plant phenotyping methodsRGB-D画像と粗い深度マップから樹幹直径を推定するモバイルアプリと深層学習手法を開発し、複数国のデータで精度・速度を検証しており、植物形質取得手法が中心である。
abstractWe present a mobile application that uses coarse-grain depth maps derived from an optical sensor
Corn, as one of the three major grain crops in China, plays a crucial role in ensuring national food security through its yield and quality. With the advancement of agricultural intelligence, agricultural robot technology has gained significant attention. High-precision navigation is the basis for realizing various operations of agricultural robots in corn fields and is closely related to the quality of operations. Corn leaf and stalk recognition and ranging are the prerequisites for achieving high-precision navigation and have attracted much attention. This paper proposes a corn leaf and stalk recognition and ranging algorithm based on multi-sensor fusion. First, YOLOv8 is used to identify corn leaves and stalks. Considering the large differences in leaf morphology and the large changes in field illumination that lead to discontinuous identification, an equidistant expansion polygon algorithm is proposed to post-process the leaves, thereby increasing the average recognition completeness of the leaves to 86.4%. Secondly, after eliminating redundant point clouds, the IMU data are used to calculate the confidence of the LiDAR and depth camera ranging point clouds, and point cloud fusion is performed based on this to achieve high-precision ranging of corn leaves. The average ranging error is 2.9 cm, which is lower than the measurement error of a single sensor. Finally, the stalk point cloud is processed and clustered using the FILL-DBSCAN algorithm to identify and measure the distance of the same corn stalk. The algorithm combines recognition accuracy and ranging accuracy to meet the needs of robot navigation or phenotypic measurement in corn fields, ensuring the stable and efficient operation of the robot in the corn field.
Why it matches plant phenotyping methodsLiDAR・深度カメラ融合によるトウモロコシ葉・茎の認識と距離計測アルゴリズムが研究の中心であり、植物器官の空間形質を抽出する方法を開発・検証している。
abstractThis paper proposes a corn leaf and stalk recognition and ranging algorithm based on multi-sensor fusion.
Potato yield is an important metric for farmers to further optimize their cultivation practices. Potato yield can be estimated on a harvester using an RGB-D camera that can estimate the three-dimensional (3D) volume of individual potato tubers. A challenge, however, is that the 3D shape derived from RGB-D images is only partially completed, underestimating the actual volume. To address this issue, we developed a 3D shape completion network, called CoRe++, which can complete the 3D shape from RGB-D images. CoRe++ is a deep learning network that consists of a convolutional encoder and a decoder. The encoder compresses RGB-D images into latent vectors that are used by the decoder to complete the 3D shape using the deep signed distance field network (DeepSDF). To evaluate our CoRe++ network, we collected partial and complete 3D point clouds of 339 potato tubers on an operational harvester in Japan. On the 1425 RGB-D images in the test set (representing 51 unique potato tubers), our network achieved a completion accuracy of 2.8 mm on average. For volumetric estimation, the root mean squared error (RMSE) was 22.6 ml, and this was better than the RMSE of the linear regression (31.1 ml) and the base model (36.9 ml). We found that the RMSE can be further reduced to 18.2 ml when performing the 3D shape completion in the center of the RGB-D image. With an average 3D shape completion time of 10 milliseconds per tuber, we can conclude that CoRe++ is both fast and accurate enough to be implemented on an operational harvester for high-throughput potato yield estimation. CoRe++'s high-throughput and accurate processing allows it to be applied to other tuber, fruit and vegetable crops, thereby enabling versatile, accurate and real-time yield monitoring in precision agriculture. Our code, network weights and dataset are publicly available at https://github.com/UTokyo-FieldPhenomics-Lab/corepp.git.
Why it matches plant phenotyping methodsRGB-D画像からジャガイモ塊茎の3D形状を補完し、体積・収量を推定する手法を開発・検証しており、植物フェノタイピング手法が研究の中心である。
abstractwe developed a 3D shape completion network, called CoRe++, which can complete the 3D shape from RGB-D images.
Reproduction assets foundThe paper's abstract explicitly states that the authors' code, network weights, and the potato tuber RGB-D/3D point cloud dataset are publicly available at the authors' GitHub repository (UTokyo-FieldPhenomics-Lab/corepp), which is a paper-specific, public, actionable asset for the CoRe++ phenotyping analysis.Code · publicOur code, network weights and dataset are publicly available at https://github.com/UTokyo-FieldPhenomics-Lab/corepp.git .Open asset ↗UTokyo-FieldPhenomics-Lab/corepplines:1-93Code / dataset availability confirmedarXiv · checked 14 Sept 2026
As the world population is expected to reach 10 billion by 2050, our agricultural production system needs to double its productivity despite a decline of human workforce in the agricultural sector. Autonomous robotic systems are one promising pathway to increase productivity by taking over labor-intensive manual tasks like fruit picking. To be effective, such systems need to monitor and interact with plants and fruits precisely, which is challenging due to the cluttered nature of agricultural environments causing, for example, strong occlusions. Thus, being able to estimate the complete 3D shapes of objects in presence of occlusions is crucial for automating operations such as fruit harvesting. In this paper, we propose the first publicly available 3D shape completion dataset for agricultural vision systems. We provide an RGB-D dataset for estimating the 3D shape of fruits. Specifically, our dataset contains RGB-D frames of single sweet peppers in lab conditions but also in a commercial greenhouse. For each fruit, we additionally collected high-precision point clouds that we use as ground truth. For acquiring the ground truth shape, we developed a measuring process that allows us to record data of real sweet pepper plants, both in the lab and in the greenhouse with high precision, and determine the shape of the sensed fruits. We release our dataset, consisting of almost 7,000 RGB-D frames belonging to more than 100 different fruits. We provide segmented RGB-D frames, with camera intrinsics to easily obtain colored point clouds, together with the corresponding high-precision, occlusion-free point clouds obtained with a high-precision laser scanner. We additionally enable evaluation of shape completion approaches on a hidden test set through a public challenge on a benchmark server.
Why it matches plant phenotyping methods果実の3D形状という植物器官の形態形質を対象に、RGB-D画像・高精度点群・評価用ベンチマークを構築しており、形状取得と推定の方法論が中心である。
abstractWe provide an RGB-D dataset for estimating the 3D shape of fruits.
Reproduction assets found保存済みの本文根拠を更新済みルールで再検証し、公開資産1件を確認しました。Code · publicOur development toolkit including a data loader is available at:
https://github.com/PRBonn/shape_completion_toolkit for handling the dataset and computing metrics.Open asset ↗PRBonn/shape_completion_toolkitlines:55-81Plant phenotyping relevance match · UnverifiedEurope PMC · checked 7 Sept 2026
The use of low-cost depth imaging sensors is investigated to automate plant pathology tests. Spatial evolution is explored to discriminate plant resistance through the hypersensitive reaction involving cotyledon loss. A high temporal frame rate and a protocol operating with batches of plants enable to compensate for the low spatial resolution of depth cameras. Despite the high density of plants, a spatial drop of the depth is observed when the cotyledon loss occurs. We introduce a small and simple spatiotemporal feature space which is shown to carry enough information to automate the discrimination between batches of resistant (loss of cotyledons) and susceptible plants (no loss of cotyledons) with 97% accuracy and with a timing 30 times faster than for human annotation. The robustness of the method-in terms of density of plants in the batch and possible internal batch desynchronization-is assessed successfully with hundreds of varieties of Pepper in various environments. A study on the generalizability of the method suggests that it can be extended to other pathosystems and also to segregating plants, i.e., intermediate state with batches composed of resistant and susceptible plants. The imaging system developed, combined with the feature extraction method and classification model, provides a full pipeline with unequaled throughput and cost efficiency by comparison with the state-of-the-art one. This system can be deployed as a decision-support tool but is also compatible with a standalone technology where computation is done at the edge in real time.
Why it matches plant phenotyping methods低コスト深度画像による植物病害表現型（子葉脱落・抵抗性）の自動抽出、分類パイプラインを開発し、多数品種・環境で性能と頑健性を検証しており、表現型取得手法が中心である。
abstractThe use of low-cost depth imaging sensors is investigated to automate plant pathology tests.
Feeding the increasing global population and reducing the carbon footprint of agricultural activities are two critical challenges of our century. Growing crops under protected horticulture and precise crop monitoring have emerged to address these challenges. Crop monitoring in commercial protected facilities remains mostly manual and labour intensive. Using computer vision to solve specific problems in image-based crop monitoring in these compact and complex growth environments is currently hindered by the scarcity of available data. We collected an RGBD dataset for vertically supported, hydroponically-grown capsicum plants in a commercial-scale glasshouse facility to fill this gap. Data were collected weekly using a single top-angled stereo camera mounted on a mobile platform running between the hydroponic gutters. The RGBD streams covered 80 % of the crop growing season in three different light conditions. The metadata include camera configurations and light condition information. Manually measured plant heights of ten selected plants per gutter are provided as ground truth. The images covered the whole plants and focused on the top third. This dataset will support research on plant height estimation, plant organ identification, object segmentation, organ measurements, 3D reconstruction, 3D data processing, and depth noise reduction. The usability of the dataset has been successfully demonstrated in a previously published study on plant height estimation using machine learning and 3D point cloud.
Why it matches plant phenotyping methods植物の草丈推定や器官計測を目的としたRGBD画像データセットを構築し、地上真値も提供しているため、植物フェノタイピング手法・データセットが中心です。
abstractWe collected an RGBD dataset for vertically supported, hydroponically-grown capsicum plants in a commercial-scale glasshouse facility to fill this gap.
Reproduction assets foundThis data article directly deposits its paper-specific phenotyping assets: the PC4C_CAPSI RGBD image dataset (Rosbag streams, JSON metadata, manual plant-height ground truth) on the Western Sydney University ResearchDirect repository, and the authors' RGBD processing code (image extraction, depth correction, 3D reconssDataset · publicData accessibility
Repository name: Image Data of Capsicum Plant Growth in Protected Horticulture: PC4C_CAPSI. [ 1 ]
Data identification number: 10.26183/1A0R-E318
Direct URL to data: https://rds.westernsydney.edu.au/Institutes/HIE/2024/Jayasuriya_N/Open asset ↗rds.westernsydney.edu.aulines:1-40Code / dataset availability confirmedOpenAlex · arXiv · checked 15 Sept 2026
The use of multiple camera technologies in a combined multimodal monitoring system for plant phenotyping offers promising benefits. Compared to configurations that only utilize a single camera technology, cross-modal patterns can be recorded that allow a more comprehensive assessment of plant phenotypes. However, the effective utilization of cross-modal patterns is dependent on precise image registration to achieve pixel-accurate alignment, a challenge often complicated by parallax and occlusion effects inherent in plant canopy imaging. In this study, we propose a novel multimodal 3D image registration method that addresses these challenges by integrating depth information from a time-of-flight camera into the registration process. By leveraging depth data, our method mitigates parallax effects and thus facilitates more accurate pixel alignment across camera modalities. Additionally, we introduce an automated mechanism to identify and differentiate different types of occlusions, thereby minimizing the introduction of registration errors. To evaluate the efficacy of our approach, we conduct experiments on a diverse image dataset comprising six distinct plant species with varying leaf geometries. Our results demonstrate the robustness of the proposed registration algorithm, showcasing its ability to achieve accurate alignment across different plant types and camera compositions. Compared to previous methods it is not reliant on detecting plant specific image features and can thereby be utilized for a wide variety of applications in plant sciences. The registration approach principally scales to arbitrary numbers of cameras with different resolutions and wavelengths. Overall, our study contributes to advancing the field of plant phenotyping by offering a robust and reliable solution for multimodal image registration.
Why it matches plant phenotyping methods植物フェノタイピングのためのマルチモーダル3D画像登録手法を開発し、複数植物種の画像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。
abstractIn this study, we propose a novel multimodal 3D image registration method that addresses these challenges by integrating depth information from a time-of-flight camera into the registration process.
Reproduction assets foundThe paper's multimodal plant image dataset (six plant species recorded with the RGBD/thermal/hyperspectral setup) is publicly available on the authors' GitHub repository, which is an allowed URL.Dataset · publiclity of our registration algorithm across diverse scenarios, we recorded a dataset comprising images of six distinct plant species. This was done to encompass a wide variety of leaf and canopy structures, thus offering a representative sample for evaluation purposes. The recorded dataset can be found on the project github page: https://github.com/eric-stumpe/Plant3DImageReg . The chosen plant species are as follows:
1.
Grapevine ( Vitis vinifera )
2.
Leopard lily ( Dieffenbachia )
3.Open asset ↗eric-stumpe/Plant3DImageReglines:287-310Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Greenhouse production of fruits and vegetables in developed countries is challenged by labour scarcity and high labour costs. Robots offer a good solution for sustainable and cost-effective production. Acquiring accurate spatial information about relevant plant parts is vital for successful robot operation. Robot perception in greenhouses is challenging due to variations in plant appearance, viewpoints, and illumination. This paper proposes a keypoint-detection-based method using data from an RGB-D camera to estimate the 3D pose of peduncle nodes, which provides essential information to harvest the tomato bunches. Specifically, this paper proposes a method that detects four anatomical landmarks in the colour image and then integrates 3D point-cloud information to determine the 3D pose. A comprehensive evaluation was conducted in a commercial greenhouse to gain insight into the performance of different parts of the method. The results showed: (1) high accuracy in object detection, achieving an Average Precision (AP) of AP@0.5=0.96; (2) an average Percentage of Detected Joints (PDJ) of the keypoints of PhDJ@0.2 = 94.31%; and (3) 3D pose estimation accuracy with mean absolute errors (MAE) of 11ᵒ and 10ᵒ for the relative upper and lower angles between the peduncle and main stem, respectively. Furthermore, the capability to handle variations in viewpoint was investigated, demonstrating the method was robust to view changes. However, canonical and higher views resulted in slightly higher performance compared to other views. Although tomato was selected as a use case, the proposed method has the potential to be applied to other greenhouse crops, such as pepper, after fine-tuning.
Why it matches plant phenotyping methodsRGB-D画像と深度点群を用いてトマトの器官ランドマークを検出し、ペドンクル節の3D姿勢を推定する手法を開発・評価しており、植物形態の取得が中心である。
abstractThis paper proposes a keypoint-detection-based method using data from an RGB-D camera to estimate the 3D pose of peduncle nodes
Introduction Yunnan Xiaomila is a pepper variety whose flowers and fruits become mature at the same time and multiple times a year. The distinction between the fruits and the background is low and the background is complex. The targets are small and difficult to identify. Methods This paper aims at the problem of target detection of Yunnan Xiaomila under complex background environment, in order to reduce the impact caused by the small color gradient changes between xiaomila and background and the unclear feature information, an improved PAE-YOLO model is proposed, which combines the EMA attention mechanism and DCNv3 deformable convolution is integrated into the YOLOv8 model, which improves the model's feature extraction capability and inference speed for Xiaomila in complex environments, and achieves a lightweight model. First, the EMA attention mechanism is combined with the C2f module in the YOLOv8 network. The C2f module can well extract local features from the input image, and the EMA attention mechanism can control the global relationship. The two complement each other, thereby enhancing the model's expression ability; Meanwhile, in the backbone network and head network, the DCNv3 convolution module is introduced, which can adaptively adjust the sampling position according to the input feature map, contributing to stronger feature capture capabilities for targets of different scales and a lightweight network. It also uses a depth camera to estimate the posture of Xiaomila, while analyzing and optimizing different occlusion situations. The effectiveness of the proposed method was verified through ablation experiments, model comparison experiments and attitude estimation experiments. Results The experimental results indicated that the model obtained an average mean accuracy (mAP) of 88.8%, which was 1.3% higher than that of the original model. Its F1 score reached 83.2, and the GFLOPs and model sizes were 7.6G and 5.7MB respectively. The F1 score ranked the best among several networks, with the model weight and gigabit floating-point operations per second (GFLOPs) being the smallest, which are 6.2% and 8.1% lower than the original model. The loss value was the lowest during training, and the convergence speed was the fastest. Meanwhile, the attitude estimation results of 102 targets showed that the orientation was correctly estimated exceed 85% of the cases, and the average error angle was 15.91°. In the occlusion condition, 86.3% of the attitude estimation error angles were less than 40°, and the average error angle was 23.19°. Discussion The results show that the improved detection model can accurately identify Xiaomila targets fruits, has higher model accuracy, less computational complexity, and can better estimate the target posture.
Why it matches plant phenotyping methods唐辛子果実の検出に加え、深度カメラによる姿勢推定手法を開発し、アブレーション・比較・姿勢推定実験で技術検証しているため、植物器官の表現型取得が中心である。
In facility horticulture smart farms, harvesting robotic systems to automate harvesting tasks are challenging due to the complex environment, irregular growth and fruits pose. To harvest a fruits, the vision information required by a harvesting robot is accurate position and pose. Especially, fruit pose information is essential for planning paths to avoid damaging stems, leaves, branches, and obstacles, preventing damage to the fruit and harvesting robots, and planning efficient harvest sequences. This paper presents a method that uses information from tomato sepals to estimate the orientation of tomatoes, which are a commonly grown crop in horticulture. First, we train a YOLOv8 model to detect and segment bodies and sepals of tomatoes. For robust training of the model, the training data is constructed using effective data augmentation, which synthesizes segmented foreground objects and inserting them into the background. Then, for finding association between the body and sepal of a tomato, we apply IoU score based matching and the Hungarian algorithm. Consequently, we obtain point clouds for both parts using RGB-D data. Finally, we compute the center point of each object by using spherical fitting and the statistics of the point cloud, respectively. Then, we estimate the pose of the tomato as a vector between two center points of the body and sepal. To accurately and practically evaluate the proposed method, we generated ground truth data using calibration patterns in tomato greenhouse. As the experimental results show, The segmentation results show that AP50,sepal is 94.7, AP50,tomato is 96.3, and mAP is 61.5. The result of the pose estimation for the pose validation dataset is a mean error angle of 6.79 ± 3.18 and angle errors of less than 10 degrees account for 87.2%. The total algorithm proposed in this paper requires about 0.038s for each tomato greenhouse image. As a result, our proposed pose estimation can be practically utilized in robot systems for tomato harvesting.
Why it matches plant phenotyping methodsトマト果実の姿勢という植物器官の形態特性を、画像・RGB-Dデータから推定する手法を開発し、精度検証まで行っているため、中心的な植物フェノタイピング研究である。
abstractThis paper presents a method that uses information from tomato sepals to estimate the orientation of tomatoes
The detection and sizing of fruits with computer vision methods is of interest because it provides relevant information to improve the management of orchard farming. However, the presence of partially occluded fruits limits the performance of existing methods, making reliable fruit sizing a challenging task. While previous fruit segmentation works limit segmentation to the visible region of fruits (known as modal segmentation), in this work we propose an amodal segmentation algorithm to predict the complete shape, which includes its visible and occluded regions. To do so, an end-to-end convolutional neural network (CNN) for simultaneous modal and amodal instance segmentation was implemented. The predicted amodal masks were used to estimate the fruit diameters in pixels. Modal masks were used to identify the visible region and measure the distance between the apples and the camera using the depth image. Finally, the fruit diameters in millimetres (mm) were computed by applying the pinhole camera model. The method was developed with a Fuji apple dataset consisting of 3925 RGB-D images acquired at different growth stages with a total of 15,335 annotated apples, and was subsequently tested in a case study to measure the diameter of Elstar apples at different growth stages. Fruit detection results showed an F1-score of 0.86 and the fruit diameter results reported a mean absolute error (MAE) of 4.5 mm and R² = 0.80 irrespective of fruit visibility. Besides the diameter estimation, modal and amodal masks were used to automatically determine the percentage of visibility of measured apples. This feature was used as a confidence value, improving the diameter estimation to MAE = 2.93 mm and R² = 0.91 when limiting the size estimation to fruits detected with a visibility higher than 60%. The main advantages of the present methodology are its robustness for measuring partially occluded fruits and the capability to determine the visibility percentage. The main limitation is that depth images were generated by means of photogrammetry methods, which limits the efficiency of data acquisition. To overcome this limitation, future works should consider the use of commercial RGB-D sensors. The code and the dataset used to evaluate the method have been made publicly available at https://github.com/GRAP-UdL-AT/Amodal_Fruit_Sizing.
Why it matches plant phenotyping methodsリンゴの遮蔽下での果実径という植物器官形質を推定する画像解析手法を開発し、データセットで検証しているため、方法が中心的である。
abstractin this work we propose an amodal segmentation algorithm to predict the complete shape, which includes its visible and occluded regions.
Field / plotRGB-D / ToFStem / branchImage / point-cloud registrationSegmentation
Automated robots are emerging as a solution for labor-intensive fruit orchard management. Three-dimensional (3D) reconstruction of tree branches is a fundamental requirement for robots to perform tasks like pruning and fruit harvesting. Current branch sensing methods often rely on planar segmentation with limited 3D information or computationally expensive point cloud segmentation, which may not be suitable for natural orchards with obscured tree branches. This study proposes a novel scheme that reconstructs occluded branches from RGB-D (Red-Green-Blue-Depth) images by integrating the point clouds converted from planar segmentation masks and depth images. The proposed approach extends the existing 2D branch sensing techniques to 3D, leveraging multi-view information. The deep learning models DeeplabV3+ and Pix2pix are employed to generate the segmentation masks, separately. And the Fast Global Registration (FGR) is used to register the multi-view point clouds. The results demonstrate that the output point clouds have at least a 24 % increase in the number of corresponding points after FGR. Furthermore, the time cost per hundred corresponding points is reduced by 85 % and 69 % when using the DeepLabV3 + and Pix2pix-based schemes, respectively, compared to the PointNet++ approach. These findings indicate that the proposed scheme significantly improves the sensing of occluded branches in terms of output richness and computational efficiency, making it applicable to natural orchard working spaces.
Why it matches plant phenotyping methodsRGB-D画像のセグメンテーションと点群登録により、遮蔽された樹枝の3D形状を再構成する植物形質センシング手法の開発が中心である。
abstractThis study proposes a novel scheme that reconstructs occluded branches from RGB-D (Red-Green-Blue-Depth) images by integrating the point clouds converted from planar segmentation masks and depth images.
LiDAR / point cloudRGB-D / ToFImage / point-cloud registration
Abstract Precision agriculture benefits from point set registration, which can monitor plant health and growth in real time, promote the precise application of fertilizers and pesticides, and provide technical support for achieving sustainable development of agriculture. In this work, we propose a robust point set registration method for precision agriculture based on L*a*b* color guidance, bidirectional search and Cauchy distribution. First, the L*a*b* color guidance is applied to establish accurate correspondences between agricultural RGB‐D data. Second, the bidirectional nearest neighbor search strategy between point sets improves the reliability of establishing correspondences and broadens the convergence domain of the algorithm. Third, Cauchy distribution is utilized as an energy function for noise suppression, which further improves the robustness of the algorithm in dealing with complex vegetation scenes. Finally, results of ablation and simulation experiments indicate that the proposed registration algorithm can achieve more accurate and robust alignment results than other classic and state‐of‐the‐art point cloud registration algorithms to achieve monitoring and comparison of plant growth.
Why it matches plant phenotyping methods植物RGB-D点群を対象に、色彩誘導・双方向探索・Cauchyカーネルを用いた点群位置合わせ手法を開発し、植物の成長モニタリングと比較に適用・評価しているため、表現型取得の中核手法に該当する。
abstractwe propose a robust point set registration method for precision agriculture based on L*a*b* color guidance, bidirectional search and Cauchy distribution.
Precise detection and localization are prerequisites for intelligent harvesting, while fruit size and weight estimation are key to intelligent orchard management. In commercial banana orchards, it is necessary to manage the growth and weight of banana bunches so that they can be harvested in time and prepared for transportation according to their different maturity levels. In this study, in order to reduce management costs and labor dependence, and obtain non-destructive weight estimation, we propose a method for localizing and estimating banana bunches using RGB-D images. First, the color image is detected through the YOLO-Banana neural network to obtain two-dimensional information about the banana bunches and stalks. Then, the three-dimensional coordinates of the central point of the banana stalk are calculated according to the depth information, and the banana bunch size is obtained based on the depth information of the central point. Finally, the effective pixel ratio of the banana bunch is presented, and the banana bunch weight estimation model is statistically analyzed. Thus, the weight estimation of the banana bunch is obtained through the bunch size and the effective pixel ratio. The R2 value between the estimated weight and the actual measured value is 0.8947, the RMSE is 1.4102 kg, and the average localization error of the central point of the banana stalk is 22.875 mm. The results show that the proposed method can provide bunch size and weight estimation for the intelligent management of banana orchards, along with localization information for banana-harvesting robots.
Why it matches plant phenotyping methodsRGB-D画像からバナナ房のサイズと重量という植物器官形質を推定する手法を開発・評価しており、収穫ロボット用の局在化にとどまらないため。
abstractwe propose a method for localizing and estimating banana bunches using RGB-D images
Field / plotLiDAR / point cloudRGB-D / ToFMultispectral / hyperspectralThermalWhole plant / canopy / plot / fieldCalibration / preprocessing2D/3D reconstruction
Abstract Non-invasive crop phenotyping is essential for crop modeling, which relies on image processing techniques. This research presents a plant-scale vision system that can acquire multispectral plant data in agricultural fields. This paper proposes a sensory fusion method that uses three cameras, Two multispectral and a RGB depth camera. The sensory fusion method applies pattern recognition and statistical optimization to produce a single multispectral 3D image that combines thermal and near-infrared (NIR) images from crops. A multi-camera sensory fusion method incorporates five multispectral bands: three from the visible range and two from the non-visible range, namely NIR and mid-infrared. The object recognition method examines about 7000 features in each image and runs only once during calibration. The outcome of the sensory fusion process is a homographic transformation model that integrates multispectral and RGB data into a coherent 3D representation. This approach can handle occlusions, allowing an accurate extraction of crop features. The result is a 3D point cloud that contains thermal and NIR multispectral data that were initially obtained separately in 2D.
Why it matches plant phenotyping methods植物スケールのマルチカメラ融合による3D・マルチスペクトルデータ取得と作物特徴抽出を開発した、植物フェノタイピング手法が中心の研究。
abstractThis research presents a plant-scale vision system that can acquire multispectral plant data in agricultural fields.
Crop yield estimation plays a crucial role in agricultural production planning and risk management. Utilizing simultaneous localization and mapping (SLAM) technology for the three-dimensional reconstruction of crops allows for an intuitive understanding of their growth status and facilitates yield estimation. Therefore, this paper proposes a VINS-RGBD system incorporating a semantic segmentation module to enrich the information representation of a 3D reconstruction map. Additionally, image matching using L_SuperPoint feature points is employed to achieve higher localization accuracy and obtain better map quality. Moreover, Voxblox is proposed for storing and representing the maps, which facilitates the storage of large-scale maps. Furthermore, yield estimation is conducted using conditional filtering and RANSAC spherical fitting. The results show that the proposed system achieves an average relative error of 10.87% in yield estimation. The semantic segmentation accuracy of the system reaches 73.2% mIoU, and it can save an average of 96.91% memory for point cloud map storage. Localization accuracy tests on public datasets demonstrate that, compared to Shi–Tomasi corner points, using L_SuperPoint feature points reduces the average ATE by 1.933 and the average RPE by 0.042. Through field experiments and evaluations in a strawberry field, the proposed system demonstrates reliability in yield estimation, providing guidance and support for agricultural production planning and risk management.
Why it matches plant phenotyping methods3D再構成・意味分割・RANSACによるイチゴ収量推定システムを開発し、精度評価まで行っており、植物形質（収量）の取得・推定法が中心である。
abstractthis paper proposes a VINS-RGBD system incorporating a semantic segmentation module to enrich the information representation of a 3D reconstruction map.
Greenhouse production of fruits and vegetables in developed countries is challenged by labour scarcity and high labour costs. Robots offer a good solution for sustainable and cost-effective production. Acquiring accurate spatial information about relevant plant parts is vital for successful robot operation. Robot perception in greenhouses is challenging due to variations in plant appearance, viewpoints, and illumination. This paper proposes a keypoint-detection-based method using data from an RGB-D camera to estimate the 3D pose of peduncle nodes, which provides essential information to harvest the tomato bunches. Specifically, this paper proposes a method that detects four anatomical landmarks in the colour image and then integrates 3D point-cloud information to determine the 3D pose. A comprehensive evaluation was conducted in a commercial greenhouse to gain insight into the performance of different parts of the method. The results showed: (1) high accuracy in object detection, achieving an Average Precision (AP) of [email protected]=0.96; (2) an average Percentage of Detected Joints (PDJ) of the keypoints of [email protected] = 94.31%; and (3) 3D pose estimation accuracy with mean absolute errors (MAE) of 11o and 10o for the relative upper and lower angles between the peduncle and main stem, respectively. Furthermore, the capability to handle variations in viewpoint was investigated, demonstrating the method was robust to view changes. However, canonical and higher views resulted in slightly higher performance compared to other views. Although tomato was selected as a use case, the proposed method has the potential to be applied to other greenhouse crops, such as pepper, after fine-tuning.
Why it matches plant phenotyping methodsトマトの器官ランドマークを画像・点群から抽出し、ペドンクル節の3D姿勢という植物形態形質を推定する手法を開発・評価しており、収穫ロボットへの応用を超えて表現型取得が中心である。
abstractThis paper proposes a keypoint-detection-based method using data from an RGB-D camera to estimate the 3D pose of peduncle nodes
Stem diameter is a critical phenotypic parameter for maize, integral to yield prediction and lodging resistance assessment. Traditionally, the quantification of this parameter through manual measurement has been the norm, notwithstanding its tedious and laborious nature. To address these challenges, this study introduces a non-invasive field-based system utilizing depth information from RGB-D cameras to measure maize stem diameter. This technology offers a practical solution for conducting rapid and non-destructive phenotyping. Firstly, RGB images, depth images, and 3D point clouds of maize stems were captured using an RGB-D camera, and precise alignment between the RGB and depth images was achieved. Subsequently, the contours of maize stems were delineated using 2D image processing techniques, followed by the extraction of the stem's skeletal structure employing a thinning-based skeletonization algorithm. Furthermore, within the areas of interest on the maize stems, horizontal lines were constructed using points on the skeletal structure, resulting in 2D pixel coordinates at the intersections of these horizontal lines with the maize stem contours. Subsequently, a back-projection transformation from 2D pixel coordinates to 3D world coordinates was achieved by combining the depth data with the camera's intrinsic parameters. The 3D world coordinates were then precisely mapped onto the 3D point cloud using rigid transformation techniques. Finally, the maize stem diameter was sensed and determined by calculating the Euclidean distance between pairs of 3D world coordinate points. The method demonstrated a Mean Absolute Percentage Error ( MAPE ) of 3.01%, a Mean Absolute Error ( MAE ) of 0.75 mm, a Root Mean Square Error ( RMSE ) of 1.07 mm, and a coefficient of determination ( R ²) of 0.96, ensuring accurate measurement of maize stem diameter. This research not only provides a new method of precise and efficient crop phenotypic analysis but also offers theoretical knowledge for the advancement of precision agriculture.
Why it matches plant phenotyping methodsRGB-Dカメラと画像・3D処理によりトウモロコシ茎径を非破壊測定する手法を開発し、誤差指標で精度検証しており、フェノタイピング手法が中心である。
abstractthis study introduces a non-invasive field-based system utilizing depth information from RGB-D cameras to measure maize stem diameter
Reproduction assets foundThe paper's data availability statement points to a public Figshare deposit (DOI 10.6084/m9.figshare.25450039) containing the study's datasets (RGB/depth imagery and stem diameter measurements used for the maize stem diameter phenotyping analysis). No author analysis code or trained models are explicitly deposited.Dataset · publicThe datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: http://dx.doi.org/10.6084/m9.figshare.25450039 .Open asset ↗figshare · 10.6084/m9.figshare.25450039lines:909-917Code / dataset availability confirmedOpenAlex · checked 15 Sept 2026
Trunk diameter is related to the overall health and level of carbon sequestration in a tree. Trunk diameter measurement, therefore, is a key task in both forest plot and urban settings. Unlike the traditional approach of manual measurement with a measuring tape or calipers, several recent approaches rely on sophisticated technologies such as Terrestrial Laser Scanning (TLS), LiDAR, and time-of-flight sensors that provide fine-grain depth maps, which are used for depth-assisted image segmentation in downstream processing. These technologies are supported only on specialized devices or high-end smartphones. We present a mobile application that uses coarse-grain depth maps derived from an optical sensor, and so can be run on most common Android devices. Moreover, we use a state-of-the-art deep neural network to estimate trunk diameter from an image and its corresponding coarse depth map (RGB-D). We tested our app using a dataset collected from four countries and under challenging conditions including occlusion, leaning trees, and irregular shapes and found that our algorithm has a MAE of 2.58 cm and an RMSE of 3.57 cm, which is comparable to accuracy from fine-grain depth maps. Moreover, diameter measurement using our app is more than 5 times faster than traditional manual surveying.
Why it matches plant phenotyping methodsRGB-D画像と粗い深度マップから樹幹直径を推定するモバイル手法を開発し、複数国のデータセットと困難条件で精度検証しており、植物形質取得が中心である。
abstractWe present a mobile application that uses coarse-grain depth maps derived from an optical sensor
Reproduction assets foundThe paper's DBH estimation evaluation dataset (154 RGB + depth tree images with metadata and ground-truth DBH) is publicly deposited on Zenodo, and the app/algorithm source code is publicly available on GitHub. Both are paper-specific, public, and actionable.Dataset · publicing the quality of the DBH estimate, includ-
197
ing non-cylindrical trunks, burl presence, degrees of leaning and occlusion, and poor lighting, as illustrated in Fig. 4. Sample images
198
from the dataset are available in Supplement 10.9, and the complete set of RGB and depth images, along with metadata, is acces-
199
sible at https://zenodo.org/records/10199711.200
In Thailand, we collected data in Bangkok’s Lumphini Park and Chulalongkorn Centenary Park. As a tropical location, Bangkok is
201
home to many tropical trees, such as rain trees (Samanea saman), banyan trees, palm trees, and coconut trees52. At Lumphini Park,
202
where most of our data came from, small forests grow next to watOpen asset ↗zenodo.org · 10199711pdf-raw-page:8 lines:1-31Code · public; Z.F. analyzed the data and led the writing of the manuscript.
352
A.H. and S.K. reviewed the manuscript and provided constructive suggestions. All authors contributed critically to the drafts and
353
gave final approval for publication.
354
8 DATA AVAILABILITY
355
The algorithm and app code are publicly available on GitHub at https://github.com/MingyueX/GreenLens, with APK available from
356
APKPure at https://apkpure.com/p/com.cleeg.greenlens. All the data for the app evaluation can be accessed at https://zenodo.org/357
records/10199711.
358
9 AUTHOR COMPETING INTERESTS
359
The authors declare no conflict of interest.
360
361
REFERENCES
362
[1] Kenneth G MacDicken. Global forest resourOpen asset ↗github.com/MingyueX/GreenLenspdf-raw-page:15 lines:1-92Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Published1 Apr 2024Computers and Electronics in Agriculture.
Traditionally, vegetable and fruit production has relied on empirical and ambiguous decisions made by human farmers. To overcome this uncertainty in agriculture, smart farm robots have been widely studied in recent years. However, measuring growth information with robots remains a challenge because of the similarity in the appearance of the target plant and those around it. In this study, we propose a smart farm robot that accurately measures the growth information of a target plant based on object detection, image fusion, and data augmentation with fused images. The proposed smart farm robot uses an end-to-end real-time deep learning-based object detector that shows state-of-the-art performances. To distinguish the target plant from other plants with a higher accuracy and improved robustness than those of existing methods, we exploited image fusion using both RGB and depth images. In particular, the data augmentation, based on the fused RGB, and depth information, contributes to the precise measurement of growth information from smart farms, regardless of the high density of vegetables and fruits in these farms. We propose and evaluate a real-time measurement system to obtain precise target-plant growth information in precision agriculture. The code and models are publicly available on Github: https://github.com/kistvision/Plant_growth_measurement.
Why it matches plant phenotyping methods対象植物の成長情報を画像融合・物体検出で測定するリアルタイム手法を開発・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。
abstractwe propose a smart farm robot that accurately measures the growth information of a target plant based on object detection, image fusion, and data augmentation with fused images.
Accurate estimation of tomato cluster yields is critical to the advancement of intelligent and unmanned greenhouses, guiding horticultural management and adjusting sales and marketing strategies. However, due to the complex natural environment and tracking stability, there are still considerable challenges for automated yield estimation to be deployed in practice. Therefore, this paper presents an improved tomato cluster counting method that combines object detection, multiple object tracking, and specific tracking region counting. To reduce background tomato misidentification, we proposed the YOLOv5-4D that fuses RGB images and depth images as input. Next, we adopted ByteTrack to track tomato clusters in continuous frames and designed a specific tracking region counting method to overcome the problem of tracked tomato cluster ID shift. In the test set, the improved YOLOv5-4D had a detection accuracy of 97.9 % and a mAP@0.5:0.95 of 0.748. Field experiments showed that the counting method achieved a statistical average counting accuracy of 95.1 % and the integrated algorithm ran at more than 40 FPS, enabling stable real-time yield estimation.
Why it matches plant phenotyping methodsRGB-D画像、物体検出、追跡、計数を統合してトマト房数・収量を推定する手法が研究の中心であり、植物器官の表現型を直接定量している。
abstractthis paper presents an improved tomato cluster counting method that combines object detection, multiple object tracking, and specific tracking region counting.
This research paper explores a comprehensive approach to advancing capsicum harvesting by integrating cutting-edge technologies. The study addresses key objectives, including capsicum detection using various YOLO algorithms, peduncle detection through YOLO segmentation models in a proposed robotic harvester, laboratory testing of cutting target point coordinates using the Real Sense D455 RGB-D camera, growth stage determination, and capsicum counting/tracking with a supervision algorithm. The investigation highlights the YOLOv8s model as the most successful for capsicum detection, achieving a remarkable mean Average Precision (mAP) of 0.967 at a 0.5 Intersection over Union (IOU) threshold. As part of the growth stage determination task, YOLOv8s achieved a satisfactory mAP of 0.614 at the same IOU threshold. Additionally, the YOLOv8s-seg model demonstrated superior performance in peduncle detection, attaining a box mAP of 0.790 and a mask mAP of 0.771. The YOLOv8s-seg model excels in peduncle detection with a box mAP of 0.790 and a mask mAP of 0.771. Laboratory experiments using the Real Sense D455 RGB-D camera showcased its capability to localize the target point with a maximum error of 8 mm longitudinally, 9 mm vertically, and 12 mm laterally. The developed tracking and counting algorithm achieve a notable counting accuracy of 94.1 % during the third harvesting cycle in the greenhouse. The Android application developed demonstrated robust performance, achieving high accuracy (Mean IoU: 0.92), precise localization (Mean Euclidean Distance: 5 pixels), responsive user interface (Touch Response Time: 150 ms), and broad compatibility across Android versions and device types, with effective error handling (Success Rate: 95 %). The study not only advances capsicum harvesting techniques but also presents practical insights for the integration of advanced technologies, paving the way for efficient robotic harvesting systems in agriculture.
Why it matches plant phenotyping methodsYOLO画像解析、成長段階分類、果実カウント・追跡を統合した植物状態・器官形質の取得手法を開発し、精度を評価しているため、収穫ロボットの単なる対象定位を超える中心的なフェノタイピング研究である。
abstractThe study addresses key objectives, including capsicum detection using various YOLO algorithms, peduncle detection through YOLO segmentation models in a proposed robotic harvester, laboratory testing of cutting target point coordinates using the Real Sense D455 RGB-D camera, growth stage determination, and capsicum counting/tracking with a supervision algorithm.
Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · OpenAlex · checked 7 Sept 2026
Automated precision weed control requires visual methods to discriminate between crops and weeds. State-of-the-art plant detection methods fail to reliably detect weeds, especially in dense and occluded scenes. In the past, using hand-crafted detection models, both color (RGB) and depth (D) data were used for plant detection in dense scenes. Remarkably, the combination of color and depth data is not widely used in current deep learning-based vision systems in agriculture. Therefore, we collected an RGB-D dataset using a stereo vision camera. The dataset contains sugar beet crops in multiple growth stages with a varying weed densities. This dataset was made publicly available and was used to evaluate two novel plant detection models, the D-model, using the depth data as the input, and the CD-model, using both the color and depth data as inputs. For ease of use, for existing 2D deep learning architectures, the depth data were transformed into a 2D image using color encoding. As a reference model, the C-model, which uses only color data as the input, was included. The limited availability of suitable training data for depth images demands the use of data augmentation and transfer learning. Using our three detection models, we studied the effectiveness of data augmentation and transfer learning for depth data transformed to 2D images. It was found that geometric data augmentation and transfer learning were equally effective for both the reference model and the novel models using the depth data. This demonstrates that combining color-encoded depth data with geometric data augmentation and transfer learning can improve the RGB-D detection model. However, when testing our detection models on the use case of volunteer potato detection in sugar beet farming, it was found that the addition of depth data did not improve plant detection at high vegetation densities.
Why it matches plant phenotyping methodsRGB-Dデータセットと植物検出モデルを開発・評価し、密集環境での植物個体の画像ベース検出を中心的に扱っているため、植物フェノタイピング手法として収録する。
abstractTherefore, we collected an RGB-D dataset using a stereo vision camera.
In the field of agriculture, measuring the leaf area is crucial for the management of crops. Various techniques exist for this measurement, ranging from direct to indirect approaches and destructive to non-destructive techniques. The non-destructive approach is favored because it preserves the plant’s integrity. Among these, several methods utilize leaf dimensions, such as width and length, to estimate leaf areas based on specific models that consider the unique shapes of leaves. Although this approach does not damage plants, it is labor-intensive, requiring manual measurements of leaf dimensions. In contrast, some indirect non-destructive techniques leveraging convolutional neural networks can predict leaf areas more swiftly and autonomously. In this paper, we propose a new direct method using 3D point clouds constructed by semantic RGB-D (Red Green Blue and Depth) images generated by a semantic segmentation neural network and RGB-D images. The key idea is that the leaf area is quantified by the count of points depicting the leaves. This method demonstrates high accuracy, with an R2 value of 0.98 and a RMSE (Root Mean Square Error) value of 3.05 cm2. Here, the neural network’s role is to segregate leaves from other plant parts to accurately measure the leaf area represented by the point clouds, rather than predicting the total leaf area of the plant. This method is direct, precise, and non-invasive to sweet pepper plants, offering easy leaf area calculation. It can be implemented on laptops for manual use or integrated into robots for automated periodic leaf area assessments. This innovative method holds promise for advancing our understanding of plant responses to environmental changes. We verified the method’s reliability and superior performance through experiments on individual leaves and whole plants.
Why it matches plant phenotyping methodsRGB-D画像と意味分割による3D点群から葉面積を直接推定する手法を開発・検証しており、植物表現型の取得が研究の中心です。
abstractIn this paper, we propose a new direct method using 3D point clouds constructed by semantic RGB-D (Red Green Blue and Depth) images generated by a semantic segmentation neural network and RGB-D images.
The scientific progress in artificial intelligence and robotics has enabled precision viticulture to pursue sustainability and improve the final yield. For instance, monitoring the canopy volume of each plant can allow the correct ripening of the bunches. In this context, this paper proposes a novel approach for the characterization of biomass volume using images acquired in a vineyard with the low-cost Azure Kinect RGB-D camera. Semantic image segmentation is implemented using three encoder–decoder deep architectures (U-Net, DeepLabV3+, and MANet) to produce accurate masks of the vine leaf structure. In a transfer learning approach, a public dataset acquired with the Intel RealSense D435 depth camera is used to train the segmentation networks. Then, a complete pipeline to estimate possible changes in biomass volume is presented. Experiments are run to analyze the biomass removed during the trimming process of grapevine plants. The best segmentation result is obtained by the U-Net architecture with ResNet50 backbone, showing an accuracy of 92.10%, although the training and test sets consist of images acquired by different cameras. However, the DeepLabV3+ network with ResNeXt50 backbone, which scores an accuracy of 90.25% on the test set, gives the best estimate of the removed biomass, requiring the shortest time for training. These outcomes prove the potential capability of this automatic approach for controlling leaf growth and ensuring sustainable viticulture practices.
Why it matches plant phenotyping methodsRGB-D画像のセマンティックセグメンテーションと点群解析により、ブドウ樹の葉構造およびバイオマス体積を推定する手法を開発・評価しており、植物形質取得が中心です。
abstractthis paper proposes a novel approach for the characterization of biomass volume using images acquired in a vineyard with the low-cost Azure Kinect RGB-D camera.
Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · OpenAlex · checked 15 Sept 2026
The increasing demand for quality, year-round food production in limited space has led to the widespread adoption of protected cropping. Effectively monitoring and maintaining crops within these facilities requires substantial labour and expertise. Traditional manual monitoring is labour intensive and time consuming. Therefore, non-destructive image-based techniques, particularly those utilising 3D structural data, have gained attention. We developed a stereo vision-based system to estimate the height of vertically supported tall plants in protected facilities, given plant height serves as a vital measure of crop growth. Our system uses a mobile platform with a top-angle view of a stereo vision depth camera for data acquisition and machine learning in its core for data analysis. First, we collected weekly RGB and depth (RGBD) streams from plant gutters in three glasshouse compartments with different light treatments. We used part of the RGB data collected to train and validate a deep learning segmentation model to detect plant tops and bases. Detected tops and bases of an image were then mapped to the generated 3D scene using the depth image of the same frame. Thresholds and 3D clustering are used respectively to remove background and eliminate outliers in top and base detection mapped to 3D space. Finally, the height of each plant was calculated using the cluster centres of the tops and bases of the plants. Manually measured heights of ten selected plants per environment were used to validate the height estimations. Similar growing patterns were observed between imaged and manually measured plant heights, which showed strong correlations of 0.87, 0.96, and 0.79 R2 scores, respectively, under unfiltered ambient light, Smart Glass film, and shifted light. These promising results demonstrate the feasibility of our proposed method for a vertically supported capsicum crop in a commercial-scale protected crop facility.
Why it matches plant phenotyping methodsステレオビジョンと機械学習を用いて植物体高を推定する手法を開発し、手動測定で検証しており、植物表現型の取得が研究の中心である。
abstractWe developed a stereo vision-based system to estimate the height of vertically supported tall plants in protected facilities
In plant phenotypic research, accurate organ segmentation and crop simulation are crucial for optimizing crop planting and increasing yield. In this study, an efficient method for soybean organ segmentation and phenotypic growth simulation was explored based on 3D reconstruction technology. Taking Dongnong252 soybean as the research object, the soybean phenotype acquisition system based on Kinect sensor was used to achieve high-precision and non-destructive acquisition of soybean canopy images during the whole growth period. First, conditional filtering and statistical filtering were used to remove noise. Then, combined with the Intrinsic Shape Signatures-Coherent Point Drift (ISS-CPD) and the Iterative Closest Point (ICP) algorithms, the multi-view three-dimensional (3D) canopy structure of soybean was reconstructed, which provided a reliable basis for plant stem and leaf segmentation. On this basis, the average accuracy of plant stem and leaf segmentation was 79.99% by using the Distance-field-based segmentation pipeline (DFSP) algorithm. Furthermore, the accurate information was provided for extracting and calculating 3D phenotypic parameters such as plant height, crown width, stem thickness, leaf length and leaf width from three scales of whole plant, stem and leaf. The calculated values were highly consistent with the measured values, with an average coefficient of determination of 0.9654 and an average percentage error of 3.4862%. Meanwhile, by analyzing the quantitative relationship between phenotypic parameters and physiological development time, the data-driven Richards growth simulation model was introduced to accurately predict the growth process of soybean plants. The coefficient of determination R² values of each phenotypic simulation model reached above 0.9357, which improved the goodness of fit of the model by 0.03 compared with the Logistic model, and its root mean square error (RMSE) ranged from 0.0020 to 0.1112. The research results indicated that this method had high accuracy and reliability in 3D reconstruction of soybean canopy, phenotype calculation, and growth simulation. It could provide quantitative indicators for soybean variety selection, planting, and management, and provide technical support and reference for high-throughput phenotype acquisition and analysis of field crops.
Why it matches plant phenotyping methodsKinectによる3D再構築、器官分割、表現型形質抽出、成長シミュレーションを一体的に開発・検証しており、植物表現型取得法が研究の中心である。
abstractthe soybean phenotype acquisition system based on Kinect sensor was used to achieve high-precision and non-destructive acquisition of soybean canopy images during the whole growth period.
The scientific progress in artificial intelligence and robotics has enabled precision viticulture to pursue sustainability and improve the final yield. For instance, monitoring the canopy volume of each plant can allow the correct ripening of the bunches. In this context, this paper proposes a novel approach for the characterization of biomass volume using images acquired in a vineyard with the low-cost Azure Kinect RGB-D camera. Semantic image segmentation is implemented using three encoder–decoder deep architectures (U-Net, DeepLabV3+, and MANet) to produce accurate masks of the vine leaf structure. In a transfer learning approach, a public dataset acquired with the Intel RealSense D435 depth camera is used to train the segmentation networks. Then, a complete pipeline to estimate possible changes in biomass volume is presented. Experiments are run to analyze the biomass removed during the trimming process of grapevine plants. The best segmentation result is obtained by the U-Net architecture with ResNet50 backbone, showing an accuracy of 92.10%, although the training and test sets consist of images acquired by different cameras. However, the DeepLabV3+ network with ResNeXt50 backbone, which scores an accuracy of 90.25% on the test set, gives the best estimate of the removed biomass, requiring the shortest time for training. These outcomes prove the potential capability of this automatic approach for controlling leaf growth and ensuring sustainable viticulture practices.
Why it matches plant phenotyping methodsブドウ樹の葉構造を画像セグメンテーションし、バイオマス体積と剪定による変化を推定する手法・パイプラインが研究の中心であるため。
abstractthis paper proposes a novel approach for the characterization of biomass volume using images acquired in a vineyard with the low-cost Azure Kinect RGB-D camera.
The Columbia root-knot nematode ( Meloidogyne chitwoodi ) is a destructive soil borne pest that can cause serious economic damage to potato tubers within infected, unfumigated fields. There are very few known sources of genetic resistance to root-knot nematodes and no released potato cultivars exhibit this trait. Literature indicates that nematode resistance introgressed from Solanum bulbocastanum is dominantly inherited across many genetic backgrounds. We generated a 32 family half-diallel progeny test population utilizing 3 root-knot nematode resistant clones (female) and 11 clones (male) with russet skin type (1,600 clones, between 25 – 60 clones per family). In 2023, 1,200 progeny were evaluated relative to 6 control varieties planted at high replication at the Washington State University Experiment Station in Othello, WA. A scale, RGB-D imaging conveyor, and index scoring system was used to record: total yield, the number of tubers per plant, tuber size distribution, aspect ratio, skin color, starch content, and defect severity from all samples within this population. The distribution of phenotypic values observed from the progeny suggest that choice of parent is a highly significant factor that influences almost all traits evaluated. The phenotyping strategy utilized by our team is inexpensive, expandable, and flexible enough to be adopted by many different types of vegetable breeding programs. Data from this experiment is being used to develop potato tuber defect classification models and assess the utility of adopting genomic selection within our potato breeding program.
Why it matches plant phenotyping methodsRGB-D撮像コンベヤとスコアリングシステムによるジャガイモ塊茎形質の大規模取得・評価が明示され、フェノタイピング戦略の拡張性も主題となっているため、方法応用・プラットフォーム研究として含める。
abstractA scale, RGB-D imaging conveyor, and index scoring system was used to record: total yield, the number of tubers per plant, tuber size distribution, aspect ratio, skin color, starch content, and defect severity from all samples within this population.
Covering over a third of all terrestrial land area, forests are crucial environments; as ecosystems, for farming, and for human leisure. However, they are challenging to access for environmental monitoring, for agricultural uses, and for search and rescue applications. To enter, aerial robots need to fly through dense vegetation, where foliage can be pushed aside, but occluded branches pose critical obstacles. Therefore, we propose pixel-wise depth regression of occluded branches using three different U-Net inspired architectures. Given RGB-D input of trees with partially occluded branches, the models estimate depth values of only the wooden parts of the tree. A large photorealistic simulation dataset comprising around 44 K images of nine different tree species is generated, on which the models are trained. Extensive evaluation and analysis of the models on this dataset is shown. To improve network generalization to real-world data, different data augmentation and transformation techniques are performed. The approaches are then also successfully demonstrated on real-world data of broadleaf trees from Swiss temperate forests and a tropical Masoala Rainforest. This work showcases the previously unexplored task of frame-by-frame pixel-based occluded branch depth reconstruction to facilitate robot traversal of forest environments.
Why it matches plant phenotyping methodsRGB-D画像から樹木枝の深度・三次元構造を推定する手法を開発し、シミュレーションデータセットと実環境で評価しており、植物器官の形態計測が技術的中心である。
abstractwe propose pixel-wise depth regression of occluded branches using three different U-Net inspired architectures.
Detecting and estimating size of apples during the early stages of growth is crucial for predicting yield, pest management, and making informed decisions related to crop-load management, harvest and post-harvest logistics, and marketing. Traditional fruit size measurement methods are laborious and time-consuming. This study employs the state-of-the-art YOLOv8 object detection and instance segmentation algorithm in conjunction with geometric shape fitting techniques on 3D point cloud data to accurately determine the size of immature green apples (or fruitlet) in a commercial orchard environment. The methodology utilized two RGB-D sensors: Intel RealSense D435i and Microsoft Azure Kinect DK. Notably, the YOLOv8 instance segmentation models exhibited proficiency in immature green apple detection, with the YOLOv8m-seg model achieving the highest AP@0.5 and AP@0.75 scores of 0.94 and 0.91, respectively. Using the ellipsoid fitting technique on images from the Azure Kinect, we achieved an RMSE of 2.35 mm, MAE of 1.66 mm, MAPE of 6.15 mm, and an R-squared value of 0.9 in estimating the size of apple fruitlets. Challenges such as partial occlusion caused some error in accurately delineating and sizing green apples using the YOLOv8-based segmentation technique, particularly in fruit clusters. In a comparison with 102 outdoor samples, the size estimation technique performed better on the images acquired with Microsoft Azure Kinect than the same with Intel Realsense D435i. This superiority is evident from the metrics: the RMSE values (2.35 mm for Azure Kinect vs. 9.65 mm for Realsense D435i), MAE values (1.66 mm for Azure Kinect vs. 7.8 mm for Realsense D435i), and the R-squared values (0.9 for Azure Kinect vs. 0.77 for Realsense D435i). This study demonstrated the feasibility of accurately sizing immature green fruit in early growth stages using the combined 3D sensing and shape-fitting technique, which shows promise for improved precision agricultural operations such as optimal crop-load management in orchards.
Why it matches plant phenotyping methodsRGB-Dセンシング、YOLOv8セグメンテーション、3D形状フィッティングを組み合わせ、リンゴ果実のサイズを推定・検証する手法が研究の中心であるため。
abstractThis study employs the state-of-the-art YOLOv8 object detection and instance segmentation algorithm in conjunction with geometric shape fitting techniques on 3D point cloud data to accurately determine the size of immature green apples
Detecting and estimating size of apples during the early stages of growth is crucial for predicting yield, pest management, and making informed decisions related to crop-load management, harvest and post-harvest logistics, and marketing. Traditional fruit size measurement methods are laborious and time-consuming. This study employs the state-of-the-art YOLOv8 object detection and instance segmentation algorithm in conjunction with geometric shape fitting techniques on 3D point cloud data to accurately determine the size of immature green apples (or fruitlet) in a commercial orchard environment. The methodology utilized two RGB-D sensors: Intel RealSense D435i and Microsoft Azure Kinect DK. Notably, the YOLOv8 instance segmentation models exhibited proficiency in immature green apple detection, with the YOLOv8m-seg model achieving the highest AP@0.5 and AP@0.75 scores of 0.94 and 0.91, respectively. Using the ellipsoid fitting technique on images from the Azure Kinect, we achieved an RMSE of 2.35 mm, MAE of 1.66 mm, MAPE of 6.15 mm, and an R-squared value of 0.9 in estimating the size of apple fruitlets. Challenges such as partial occlusion caused some error in accurately delineating and sizing green apples using the YOLOv8-based segmentation technique, particularly in fruit clusters. In a comparison with 102 outdoor samples, the size estimation technique performed better on the images acquired with Microsoft Azure Kinect than the same with Intel Realsense D435i. This superiority is evident from the metrics: the RMSE values (2.35 mm for Azure Kinect vs. 9.65 mm for Realsense D435i), MAE values (1.66 mm for Azure Kinect vs. 7.8 mm for Realsense D435i), and the R-squared values (0.9 for Azure Kinect vs. 0.77 for Realsense D435i). This study demonstrated the feasibility of accurately sizing immature green fruit in early growth stages using the combined 3D sensing and shape-fitting technique, which shows promise for improved precision agricultural operations such as optimal crop-load management in orchards.
Why it matches plant phenotyping methodsRGB-D画像、インスタンスセグメンテーション、3D点群の形状フィッティングにより、リンゴ果実のサイズを推定・検証する方法が研究の中心であるため。
abstractThis study employs the state-of-the-art YOLOv8 object detection and instance segmentation algorithm in conjunction with geometric shape fitting techniques on 3D point cloud data to accurately determine the size of immature green apples
Highlights Deep learning-based instance segmentation models were applied and evaluated for tomato fruit detection. Mask R-CNN with vision transformer backbone showed the highest accuracy for tomato instance detection. Size and weight estimation indexes were calculated using tomato region depth data from instance segmentation models. Area-based index has higher accuracy for weight estimation than indexes based on weight and height information. Abstract. The size and weight of fruits are crucial factors in yield prediction and determining harvesting time. Machine vision, including fruit detection, is a key technology in the automated monitoring and harvesting of fruits. In particular, deep learning-based fruit-detection methods have been actively applied. Estimation of fruit size after fruit detection requires depth information, which can be acquired using depth imaging. RGB-D cameras include color and depth information required for fruit detection and size estimation. In this study, the RGB-D imaging technique was used to estimate the size and weight of tomatoes. Furthermore, deep learning-based instance segmentation models, including Mask R-CNN, YOLACT, and RTMDet for tomato fruit detection, were trained and evaluated. The proposed method estimated the fruit width with a root mean square error (RMSE) of 4 mm, a mean absolute percentage error (MAPE) of 4.28%, and a fruit height with an RMSE of 5.12 mm and a MAPE of 6.42%. Furthermore, the weight-prediction model based on the area index estimated the tomato fruit weight with an RMSE of 19.69 g and MAPE of 9.44%. Thus, the method can be used for accurate size and weight estimation and can be applied in growth monitoring and automated tomatoes harvesting. Keywords: Deep learning, Fruit sizing, Instance segmentation, RGB-D, Tomato.
Why it matches plant phenotyping methodsRGB-D画像と深層学習によるトマト果実の検出、サイズ・重量推定手法が研究の中心であり、植物器官の形態・重量形質を定量化して技術性能も評価している。
abstractIn this study, the RGB-D imaging technique was used to estimate the size and weight of tomatoes.
High-throughput phenotyping (HTP) has become a powerful tool for gaining insights into the genetic and environmental factors that affect cotton ( Gossypium spp.) growth and yield. With the recent advances in the field of computer vision, namely the integration of deep learning algorithms, the accuracy and efficiency of HTP systems have improved dramatically, enabling them to automatically quantify such fundamental phenotypic traits as fruit identification and enumeration. However, there is currently no HTP system available for counting all the reproductive phases of cotton crop that can be deployed in agronomic field conditions throughout the growing season. This study presents CottonSense, an advanced HTP system that overcomes the challenges of deployment across multiple growth periods by effectively segmenting and enumerating cotton fruits at four stages of growth, including square, flower, closed boll, and open boll. Consequently, CottonSense enhances agronomic management through increased opportunities for data collection and analysis. Using RGB-D cameras, it captures and processes both two and three-dimensional data, facilitating a wider range of phenotypic trait extractions such as crop biomass and plant architecture. To segment the cotton fruits, a Mask-RCNN model is trained and optimized for faster inference using TensorRT. The model yields an average AP score of 79% in segmentation across the four fruit categories. Moreover, the model's accuracy in estimating total fruit count per image is validated by a strong agreement with the counts given by ten domain experts, as reflected by an R 2 value of 0.94. Furthermore, to accurately count the segmented fruits over large populations of plants, an enumeration algorithm based on a tracking strategy is developed that achieves an R 2 value of 0.93 when compared to hand-counted fruits in the field. The proposed HTP system, which is implemented entirely on an edge computing device, is cost-effective and power-efficient, making it an effective tool for high-yield cotton breeding and crop improvement. The code for CottonSense is publicly available at https://github.com/FeriBolour/CottonSense .
Why it matches plant phenotyping methods綿花の果実を画像から分割・列挙し、専門家および手作業カウントで検証する高スループット表現型解析システムの開発が中心である。
abstractThis study presents CottonSense, an advanced HTP system that overcomes the challenges of deployment across multiple growth periods by effectively segmenting and enumerating cotton fruits at four stages of growth, including square, flower, closed boll, and open boll.
Detecting and estimating size of apples during the early stages of growth is crucial for predicting yield, pest management, and making informed decisions related to crop-load management, harvest and post-harvest logistics, and marketing. Traditional fruit size measurement methods are laborious and time-consuming. This study employs the state-of-the-art YOLOv8 object detection and instance segmentation algorithm in conjunction with geometric shape fitting techniques on 3D point cloud data to accurately determine the size of immature green apples (or fruitlet) in a commercial orchard environment. The methodology utilized two RGB-D sensors: Intel RealSense D435i and Microsoft Azure Kinect DK. Notably, the YOLOv8 instance segmentation models exhibited proficiency in immature green apple detection, with the YOLOv8m-seg model achieving the highest AP@0.5 and AP@0.75 scores of 0.94 and 0.91, respectively. Using the ellipsoid fitting technique on images from the Azure Kinect, we achieved an RMSE of 2.35 mm, MAE of 1.66 mm, MAPE of 6.15 mm, and an R-squared value of 0.9 in estimating the size of apple fruitlets. Challenges such as partial occlusion caused some error in accurately delineating and sizing green apples using the YOLOv8-based segmentation technique, particularly in fruit clusters. In a comparison with 102 outdoor samples, the size estimation technique performed better on the images acquired with Microsoft Azure Kinect than the same with Intel Realsense D435i. This superiority is evident from the metrics: the RMSE values (2.35 mm for Azure Kinect vs. 9.65 mm for Realsense D435i), MAE values (1.66 mm for Azure Kinect vs. 7.8 mm for Realsense D435i), and the R-squared values (0.9 for Azure Kinect vs. 0.77 for Realsense D435i). This study demonstrated the feasibility of accurately sizing immature green fruit in early growth stages using the combined 3D sensing and shape-fitting technique, which shows promise for improved precision agricultural operations such as optimal crop-load management in orchards.
Why it matches plant phenotyping methodsYOLOv8によるリンゴ果実の検出・セグメンテーションと、RGB-D点群および楕円体フィッティングによる果実サイズ推定が研究の中心であり、技術性能の検証とセンサー比較も実施している。
abstractThis study employs the state-of-the-art YOLOv8 object detection and instance segmentation algorithm in conjunction with geometric shape fitting techniques on 3D point cloud data to accurately determine the size of immature green apples (or fruitlet) in a commercial orchard environment.
The present dataset comprises a collection of RGB-D apple tree images that can be used to train and test computer vision-based fruit detection and sizing methods. This dataset encompasses two distinct sets of data obtained from a Fuji and an Elstar apple orchards. The Fuji apple orchard sub-set consists of 3925 RGB-D images containing a total of 15,335 apples annotated with both modal and amodal apple segmentation masks. Modal masks denote the visible portions of the apples, whereas amodal masks encompass both visible and occluded apple regions. Notably, this dataset is the first public resource to incorporate on-tree fruit amodal masks. This pioneering inclusion addresses a critical gap in existing datasets, enabling the development of robust automatic fruit sizing methods and accurate fruit visibility estimation, particularly in the presence of partial occlusions. Besides the fruit segmentation masks, the dataset also includes the fruit size (calliper) ground truth for each annotated apple. The second sub-set comprises 2731 RGB-D images capturing five Elstar apple trees at four distinct growth stages. This sub-set includes mean diameter information for each tree at every growth stage and serves as a valuable resource for evaluating fruit sizing methods trained with the first sub-set. The present data was employed in the research paper titled "Looking behind occlusions: a study on amodal segmentation for robust on-tree apple fruit size estimation" [1].
Why it matches plant phenotyping methodsリンゴ果実のRGB-D画像、アノテーション、サイズ正解値を含む公開データセットで、果実サイズ推定法の開発・評価を直接支援するため、植物フェノタイピング手法のデータ資源として中心的です。
abstractenabling the development of robust automatic fruit sizing methods and accurate fruit visibility estimation
Reproduction assets foundThe article is a Data in Brief describing the AmodalAppleSize_RGB-D dataset (RGB-D apple tree images with modal/amodal segmentation masks and fruit size ground truth), publicly deposited in Dataverse (CORA) with DOI 10.34810/data916 and a direct URL. This is the paper's own phenotyping data (images, annotations, callipDataset · publicData accessibility
Repository name: Dataverse
Data identification number: https://doi.org/10.34810/data916 [2]
Direct URL to data: https://dataverse.csuc.cat/dataset.xhtml?persistentId=doi:10.34810/data916Open asset ↗Dataverse · doi:10.34810/data916lines:43-67Plant phenotyping relevance match · UnverifiedEurope PMC · checked 7 Sept 2026
Target detection technology and variable-rate spraying technology are key technologies for achieving precise and efficient pesticide application. To address the issues of low efficiency and high working environment requirements in detecting tree information during variable spraying in orchards, this study has designed a variable spraying control system. The system employed a Kinect sensor to real-time detect the canopy volume of citrus trees and adjusted the duty cycle of solenoid valves by pulse width modulation to control the pesticide application. A canopy volume calculation method was proposed, and precision tests for volume detection were conducted, with a maximum relative error of 10.54% compared to manual measurements. A nozzle flow model was designed to determine the spray decision coefficient. When the duty cycle ranged from 30% to 90%, the correlation coefficient of the flow model exceeded 0.95, and the actual flow rate of the system was similar to the theoretical flow rate. Field experiments were conducted to evaluate the spraying effectiveness of the variable spraying control system based on the Kinect sensor. The experimental results indicated that the variable spraying control system demonstrated good consistency between the theoretical spray volume and the actual spray volume. In deposition tests, compared to constant-rate spraying, the droplets under the variable-rate mode based on canopy volume exhibited higher deposition density. Although the amount of droplet deposit and coverage slightly decreased, they still met the requirements for spraying operation quality. Additionally, the variable-rate spray mode achieved the goal of reducing pesticide use, with a maximum pesticide saving rate of 57.14%. This study demonstrates the feasibility of the Kinect sensor in guiding spraying operations and provides a reference for their application in plant protection operations.
Why it matches plant phenotyping methodsKinectセンサーによる柑橘樹冠体積の推定法を開発・精度検証し、可変散布制御へ応用しており、植物形態計測が中心的な技術貢献である。
abstractThe system employed a Kinect sensor to real-time detect the canopy volume of citrus trees
Microgreens are the first leafy seedlings of edible plants. Microgreen farming is yet to be automated; the main challenge for automation is the lack of a sensory mechanism to detect and quantify microgreen phenotypes. This paper presents a novel automated microgreen phenotyping method targeting yield estimation. The paper demonstrates that phenotyping can be effectively performed using a consumer-grade RGB-D camera. First, the depth and RGB images are captured. Thereafter, the plant segments are filtered and the canopy is identified. Using image processing, the canopy height and density are calculated. Both yield prediction regression analysis and a TensorFlow learning algorithm are evaluated to estimate the yield as a function of height and canopy density. The authors believe the algorithm discussed in this paper is the first phenotyping algorithm combining RGB and depth data for microgreen yield estimation.
Why it matches plant phenotyping methodsRGB-Dカメラと画像処理・学習アルゴリズムにより、マイクログリーンの高さ・密度から収量を推定するフェノタイピング手法が研究の中心である。
abstractThis paper presents a novel automated microgreen phenotyping method targeting yield estimation.
Considering the rising concern over climate change and the need for local food security, productive blue-green roofs (PBGR) can be an effective solution to mitigate many relevant environmental issues. However, their cost of operation is high because they are intensive, and an economical operation and maintenance approach will render them as more viable alternative. Low-cost sensors with the Internet of Things can provide reliable solutions to the real-time management and distributed monitoring of such roofs through monitoring the plant as well soil conditions. This research assesses the extent to which a low-cost image sensor can be deployed to perform continuous, automated monitoring of a urban rooftop farm as a PBGR and evaluates the thermal performance of the roof for additional crops. An RGB-depth image sensor was used in this study to monitor crop growth. Images collected from weekly scans were processed by segmentation to estimate the plant heights of three crops species. The devised technique performed well for leafy and tall stem plants like okra, and the correlation between the estimated and observed growth characteristics was acceptable. For smaller plants, bright light and shadow considerably influenced the image quality, decreasing the precision. Six other crop species were monitored using a wireless sensor network to investigate how different crop varieties respond in terms of thermal performance. Celery, snow peas, and potato were measured with maximum daily cooling records, while beet and zucchini showed sound cooling effects in terms of mean daily cooling.
Why it matches plant phenotyping methodsRGB-D画像センサーとセグメンテーションにより作物の草丈を自動推定し、観測値と相関検証しており、植物表現型取得手法が中心的である。
abstractThis research assesses the extent to which a low-cost image sensor can be deployed to perform continuous, automated monitoring of a urban rooftop farm as a PBGR
Detecting and estimating size of apples during the early stages of growth is crucial for predicting yield, pest management, and making informed decisions related to crop-load management, harvest and post-harvest logistics, and marketing. Traditional fruit size measurement methods are laborious and timeconsuming. This study employs the state-of-the-art YOLOv8 object detection and instance segmentation algorithm in conjunction with geometric shape fitting techniques on 3D point cloud data to accurately determine the size of immature green apples (or fruitlet) in a commercial orchard environment. The methodology utilized two RGB-D sensors: Intel RealSense D435i and Microsoft Azure Kinect DK. Notably, the YOLOv8 instance segmentation models exhibited proficiency in immature green apple detection, with the YOLOv8m-seg model achieving the highest AP@0.5 and AP@0.75 scores of 0.94 and 0.91, respectively. Using the ellipsoid fitting technique on images from the Azure Kinect, we achieved an RMSE of 2.35 mm, MAE of 1.66 mm, MAPE of 6.15 mm, and an R-squared value of 0.9 in estimating the size of apple fruitlets. Challenges such as partial occlusion caused some error in accurately delineating and sizing green apples using the YOLOv8-based segmentation technique, particularly in fruit clusters. In a comparison with 102 outdoor samples, the size estimation technique performed better on the images acquired with Microsoft Azure Kinect than the same with Intel Realsense D435i. This superiority is evident from the metrics: the RMSE values (2.35 mm for Azure Kinect vs. 9.65 mm for Realsense D435i), MAE values (1.66 mm for Azure Kinect vs. 7.8 mm for Realsense D435i), and the R-squared values (0.9 for Azure Kinect vs. 0.77 for Realsense D435i).
Why it matches plant phenotyping methodsYOLOv8による検出・セグメンテーションと3D形状フィッティングを組み合わせ、リンゴ果実のサイズという植物器官形質を推定・検証する手法が研究の中心である。
abstractThis study employs the state-of-the-art YOLOv8 object detection and instance segmentation algorithm in conjunction with geometric shape fitting techniques on 3D point cloud data to accurately determine the size of immature green apples
Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · OpenAlex · checked 15 Sept 2026
The inner leaves of crop canopies are obscured by outer branches and leaves, leading to information loss and observation difficulty of the occluded canopy structure using modern crop monitoring techniques. It has restricted the development of phenotypic analysis and precision agriculture. In this paper, we propose a neural network approach to reconstruct the occluded structure of crop canopies with an RGB-D sensor. Taking the cotton plant as the object of study, we propose a novel Cascade Leaf Segmentation and Completion Network (CLSCN) to reconstruct the occluded leaf images and propose a Fragmental Leaf Point–cloud Reconstruction Algorithm (FLPRA) to complete the missing point clouds. By combining the Instance Segmentation Network (ISN), Generative Adversarial Network (GAN) and Point-cloud Reconstruction Algorithm (PRA), the three-dimensional models of cotton plants with both completed internal and external structures of the canopy are smoothly reconstructed. Firstly, we collect a large number of leaf images and point clouds of cotton plants using an RGB-D sensor with the top view and construct a manually labeled cotton leaf dataset for training and evaluation. Secondly, a network named CLSCN is cascading constructed with an Instance Segmentation Network (ISN) and a Generative Adversarial Network (GAN), and the two parts of CLSCN are separately trained with our constructed dataset to output complete cotton leaves. Thirdly, with the fusion of the completed RGB images output by cascaded network segmentation and the point clouds captured by RGB-D sensor, the proposed FLPRA is used to filter, reconstruct, fuse and register the cotton canopy leaf point clouds, and to obtain the whole cotton canopy point-clouds with inner occluded structure recovery. Finally, the CLSCN and FLPRA are validated using the validation dataset of cotton leaf. The test results indicate that the front-end ISN of the proposed CLSCN can generate high-quality cotton leaf masks, with FID scores less than 35 and mIoU up to 84.65%. Additionally, the back-end GAN of CLSCN can complete the occluded leaves with an accuracy of over 94%. The reconstruction accuracy of the final three-dimensional model of the cotton canopy is as high as 82.70%. Therefore, the proposed neural network and algorithm effectively solve the problem of incomplete canopy point cloud caused by the occlusion of outer leaves and provide an effective way to recover the complete three-dimensional structure of crop canopy with internal occlusion. It is a meaningful theoretical and technical support to realize real-time crop status observation and precise field management in agriculture production.
Why it matches plant phenotyping methodsRGB-D画像、ニューラルネットワーク、点群再構成を用いて、遮蔽されたワタ個体群の3次元構造を復元する手法を開発し、データセットと精度検証も行っており、植物表現型取得が中心である。
abstractwe propose a neural network approach to reconstruct the occluded structure of crop canopies with an RGB-D sensor.
Rapid and precise tree Diameter at Breast Height (DBH) measurement is pivotal in forest inventories. While the recent advancements in LiDAR and Structure from Motion (SFM) technologies have paved the way for automated DBH measurements, the significant equipment costs and the complexity of operational procedures continue to constrain the ubiquitous adoption of these technologies for real-time DBH assessments. In this research, we introduce KAN-Forest, a real-time DBH measurement and key point localization algorithm utilizing RGB-D (Red, Green, Blue-Depth) imaging technology. Firstly, we improved the YOLOv5-seg segmentation module with a Channel and Spatial Attention (CBAM) module, augmenting its efficiency in extracting the tree’s edge features in intricate forest scenarios. Subsequently, we devised an image processing algorithm for real-time key point localization and DBH measurement, leveraging historical data to fine-tune current frame assessments. This system facilitates real-time image data upload via wireless LAN for immediate host computer processing. We validated our approach on seven sample plots, achieving bbAP50 and segAP50 scores of: 90.0%(+3.0%), 90.9%(+0.9%), respectively with the improved YOLOv5-seg model. The method exhibited a DBH estimation RMSE of 17.61∼54.96 mm (R2=0.937), and secured 78% valid DBH samples at a 59 FPS. Our system stands as a cost-effective, portable, and user-friendly alternative to conventional forest survey techniques, maintaining accuracy in real-time measurements compared to SFM- and LiDAR-based algorithms. The integration of WLAN and its inherent scalability facilitates deployment on Unmanned Ground Vehicles (UGVs) to improve the efficiency of forest inventory. We have shared the algorithms and datasets on Github for peer evaluations.
Why it matches plant phenotyping methodsRGB-D画像とアルゴリズムを用いて樹木のDBHという明示的な形態形質をリアルタイム推定し、精度検証とシステム実装を行った研究であり、植物フェノタイピング手法が中心である。
abstractwe introduce KAN-Forest, a real-time DBH measurement and key point localization algorithm utilizing RGB-D (Red, Green, Blue-Depth) imaging technology.
Reproduction assets foundThe authors explicitly state that the code used in this DBH measurement research is publicly available on GitHub (KAN-Forest repository), making it a paper-specific, public, actionable code asset. The phenotype/trait datasets (DBH measurements and forest images) are only available upon request from the corresponding作者,Code · publicwe have made
the code used in this research available on GitHub at: https://github.com/CharmingZh/KAN-ForestOpen asset ↗CharmingZh/KAN-Forestpdf-page:27 lines:1-59Plant phenotyping relevance match · UnverifiedEurope PMC · OpenAlex · checked 15 Sept 2026
The utilization of 3-dimensional point cloud technology for non-invasive measurement of plant phenotypic parameters can furnish important data for plant breeding, agricultural production, and diverse research applications. Nevertheless, the utilization of depth sensors and other tools for capturing plant point clouds often results in missing and incomplete data due to the limitations of 2.5D imaging features and leaf occlusion. This drawback obstructed the accurate extraction of phenotypic parameters. Hence, this study presented a solution for incomplete flowering Chinese Cabbage point clouds using Point Fractal Network-based techniques. The study performed experiments on flowering Chinese Cabbage by constructing a point cloud dataset of their leaves and training the network. The findings demonstrated that our network is stable and robust, as it can effectively complete diverse leaf point cloud morphologies, missing ratios, and multi-missing scenarios. A novel framework is presented for 3D plant reconstruction using a single-view RGB-D (Red, Green, Blue and Depth) image. This method leveraged deep learning to complete localized incomplete leaf point clouds acquired by RGB-D cameras under occlusion conditions. Additionally, the extracted leaf area parameters, based on triangular mesh, were compared with the measured values. The outcomes revealed that prior to the point cloud completion, the R 2 value of the flowering Chinese Cabbage's estimated leaf area (in comparison to the standard reference value) was 0.9162. The root mean square error (RMSE) was 15.88 cm 2 , and the average relative error was 22.11%. However, post-completion, the estimated value of leaf area witnessed a significant improvement, with an R 2 of 0.9637, an RMSE of 6.79 cm 2 , and average relative error of 8.82%. The accuracy of estimating the phenotypic parameters has been enhanced significantly, enabling efficient retrieval of such parameters. This development offers a fresh perspective for non-destructive identification of plant phenotypes.
Why it matches plant phenotyping methods深層学習による葉の点群補完と3D再構成を開発し、葉面積推定を比較検証しており、植物表現型取得法が研究の中心である。
abstractThis method leveraged deep learning to complete localized incomplete leaf point clouds acquired by RGB-D cameras under occlusion conditions.
Data acquired using an RGB-D Azure Kinect DK camera were used to assess different automatic algorithms to estimate the size, and predict the weight of non-occluded and occluded apples. The programming of the algorithms included: (i) the extraction of images of regions of interest (ROI) using manual delimitation of bounding boxes or binary masks; (ii) estimating the lengths of the major and minor geometric axes for the purpose of apple sizing; and (iii) predicting the final weight by allometric modelling. In addition to the use of bounding boxes, the algorithms also allowed other post-mask settings (circles, ellipses and rotated rectangles) to be implemented, and different depth options (distance between the RGB-D camera and the fruits detected) for subsequent sizing through the application of the thin lens theory. Both linear and nonlinear allometric models demonstrated the ability to predict apple weight with a high degree of accuracy (R² greater than 0.942 and RMSE < 16 g). With respect to non-occluded apples, the best weight predictions were achieved using a linear allometric model including both the major and minor axes of the apples as predictors. The mean absolute percentage error (MAPE) ranged from 5.1% to 5.7% with respective RMSE of 11.09 g and 13.02 g, depending to whether circles, ellipses, or bounding boxes were used to adjust fruit shape. The results were therefore promising and open up the possibility of implementing reliable in-field apple measurements in real time. Importantly, final weight prediction error and intermediate size estimation errors (from sizing algorithms) interact but in a way that is not easily quantifiable when weight allometric models with implicit prediction error are used. In addition, allometric models should be reviewed when applied to other apple cultivars, fruit development stages or even for different fruit growth conditions depending on canopy management.
Why it matches plant phenotyping methodsRGB-D画像からリンゴ果実のサイズと重量を推定する自動処理アルゴリズムを開発・評価しており、植物器官形質の取得が研究の中心である。
abstractassess different automatic algorithms to estimate the size, and predict the weight of non-occluded and occluded apples
과일의 품질은 색택, 크기, 모양, 결함, 당도 등 여러 요인에 의하여 결정된다. 하지만 과실의 크기, 모양 등 기하학적 정보에 대한 선별은 대부분의 농산물 산지유통센터에서 수작업으로 이루어지고 있다. 과실의 기하학적 정보의 처리를 자동화하기 위하여 2차원 영상 기반의 연구들이 수행되었지만, 한 면에 대한 2차원 영상으로 얻을 수 있는 정보는 제한적이다. 3차원 영상 정보를 농산물에 이용하고자 하는 선행 연구들이 수행되었지만, 복잡한 처리 과정으로 인해 수확 후 선별 라인에서 적용 가능한 실시간의 처리가 불가능하였다. 본 연구에서는 이러한 문제점을 해결하고 과일의 3차원 기하학 정보를 정확히 측정하고자 실시간 3차원 영상 측정 기술을 개발하고, 배 과실에 대하여 부피 계측 정확도를 평가하였다. 여러 개의 RGBD 카메라를 이용하여 여러 면의 영상 촬영이 동시에 이루어지도록 하였으며, 카메라 간의 위치 보정을 사전에 수행하여 특징점 매칭 등 시간 소모적인 작업이 촬영 시마다 수행되지 않도록 하였다. 또한, 영상 및 점군 처리를 통하여 정확한 부피 계측이 가능한 알고리즘을 설계하였다. 그 결과, 결정계수 0.9931, 평균 절대 백분율 오차 0.61%의 성능으로 배 과실의 부피가 계측 가능함을 확인하였다.
Why it matches plant phenotyping methods배 과실의 3차원 기하학적 형질인 부피를 RGB-D 다중 카메라와 점군 처리로 자동 측정하는 기술을 개발하고 정확도를 평가한 방법론 중심 연구이다.
abstract본 연구에서는 이러한 문제점을 해결하고 과일의 3차원 기하학 정보를 정확히 측정하고자 실시간 3차원 영상 측정 기술을 개발하고, 배 과실에 대하여 부피 계측 정확도를 평가하였다.
Computer vision provides a real-time, non-destructive, and indirect way of horticultural crop yield estimation. Deep learning helps improve horticultural crop yield estimation accuracy. However, the accuracy of current estimation models based on RGB (red, green, blue) images does not meet the standard of a soft sensor. Through enriching more data and improving the RGB estimation model structure of convolutional neural networks (CNNs), this paper increased the coefficient of determination (R2) by 0.0284 and decreased the normalized root mean squared error (NRMSE) by 0.0575. After introducing a novel loss function mean squared percentage error (MSPE) that emphasizes the mean absolute percentage error (MAPE), the MAPE decreased by 7.58%. This paper develops a lettuce fresh weight estimation method through the multi-modal fusion of RGB and depth (RGB-D) images. With the multimodal fusion based on calibrated RGB and depth images, R2 increased by 0.0221, NRMSE decreased by 0.0427, and MAPE decreased by 3.99%. With the novel loss function, MAPE further decreased by 1.27%. A MAPE of 8.47% helps to develop a soft sensor for lettuce fresh weight estimation.
Why it matches plant phenotyping methodsRGB-D画像と深層学習を用いてレタス生体重を推定する手法を開発し、精度指標で検証しているため、植物表現型取得が研究の中心です。
abstractThis paper develops a lettuce fresh weight estimation method through the multi-modal fusion of RGB and depth (RGB-D) images.
Photogrammetry is the science of obtaining a 3D scan of an object. Through this process, reliable information about the physical object's complex structure can be obtained, studied and analysed. A low-cost Structure from Motion (SfM) technique can be used to create 3D models using multiple 2D images from different viewpoints. A point cloud is a widely used 3D data form, which can be produced by depth sensors, such as LIDARs and RGB-D cameras. However, the cost of such scanners can be prohibitive, putting photogrammetry out of reach for many researchers and practitioners in the agriculture industry. We are developing a low-cost close-range photogrammetry rig that could be a beneficial tool for agronomists, plant scientists, and breeders. Our imaging system utilizes the Raspberry Pi to capture images with multiple cameras, and a commercial rotatory table to get images from different viewpoints. We discuss the development of extracting quantitative trait indices in wheat in order to automatically characterize planophile versus erectophile canopy architectures. Moving forward, we plan to use our photogrammetry rig for a variety of applications such as growth monitoring and extracting plant traits such as number of leaves, stem height, leaf length, leaf width, leaf area, and canopy volume. We also plan on developing bespoke, plug-and-play systems that are tailored to the specific needs of a researcher and can be operated with minimal expertise.
Why it matches plant phenotyping methods低コスト多視点フォトグラメトリ装置を開発し、コムギの草冠構造を定量化する画像ベース表現型計測が中心である。
abstractWe are developing a low-cost close-range photogrammetry rig that could be a beneficial tool for agronomists, plant scientists, and breeders.
Photogrammetry is the science of obtaining a 3D scan of an object. Through this process, reliable information about the physical object’s complex structure can be obtained, studied and analysed. A low-cost Structure from Motion (SfM) technique can be used to create 3D models using multiple 2D images from different viewpoints. A point cloud is a widely used 3D data form, which can be produced by depth sensors, such as LIDARs and RGB-D cameras. However, the cost of such scanners can be prohibitive, putting photogrammetry out of reach for many researchers and practitioners in the agriculture industry. We are developing a low-cost close-range photogrammetry rig that could be a beneficial tool for agronomists, plant scientists, and breeders. Our imaging system utilizes the Raspberry Pi to capture images with multiple cameras, and a commercial rotatory table to get images from different viewpoints. We discuss the development of extracting quantitative trait indices in wheat in order to automatically characterize planophile versus erectophile canopy architectures. Moving forward, we plan to use our photogrammetry rig for a variety of applications such as growth monitoring and extracting plant traits such as number of leaves, stem height, leaf length, leaf width, leaf area, and canopy volume. We also plan on developing bespoke, plug-and-play systems that are tailored to the specific needs of a researcher and can be operated with minimal expertise.
Why it matches plant phenotyping methods低コストの多視点画像撮影・SfM装置を開発し、コムギのキャノピー構造を定量化する方法が中心であるため。
abstractWe are developing a low-cost close-range photogrammetry rig that could be a beneficial tool for agronomists, plant scientists, and breeders.
Data acquired using an RGB-D Azure Kinect DK camera were used to assess different automatic algorithms to estimate the size, and predict the weight of non-occluded and occluded apples. The programming of the algorithms included: (i) the extraction of images of regions of interest (ROI) using manual delimitation of bounding boxes or binary masks; (ii) estimating the lengths of the major and minor geometric axes for the purpose of apple sizing; and (iii) predicting the final weight by allometric modelling. In addition to the use of bounding boxes, the algorithms also allowed other post-mask settings (circles, ellipses and rotated rectangles) to be implemented, and different depth options (distance between the RGB-D camera and the fruits detected) for subsequent sizing through the application of the thin lens theory. Both linear and nonlinear allometric models demonstrated the ability to predict apple weight with a high degree of accuracy (R2 greater than 0.942 and RMSE < 16 g). With respect to non-occluded apples, the best weight predictions were achieved using a linear allometric model including both the major and minor axes of the apples as predictors. The mean absolute percentage error (MAPE) ranged from 5.1% to 5.7% with respective RMSE of 11.09 g and 13.02 g, depending to whether circles, ellipses, or bounding boxes were used to adjust fruit shape. The results were therefore promising and open up the possibility of implementing reliable in-field apple measurements in real time. Importantly, final weight prediction error and intermediate size estimation errors (from sizing algorithms) interact but in a way that is not easily quantifiable when weight allometric models with implicit prediction error are used. In addition, allometric models should be reviewed when applied to other apple cultivars, fruit development stages or even for different fruit growth conditions depending on canopy management.
Why it matches plant phenotyping methodsRGB-D画像と自動処理アルゴリズムを用いてリンゴのサイズ・重量を推定し、複数アルゴリズムとモデルの精度を評価しているため、植物表現型取得法が中心である。
titleAssessing automatic data processing algorithms for RGB-D cameras to predict fruit size and weight in apples
Maize ( Zea mays L.) is one of the most important crops, influencing food production and even the whole industry. In recent years, global crop production has been facing great challenges from diseases. However, most of the traditional methods make it difficult to efficiently identify disease-related phenotypes in germplasm resources, especially in actual field environments. To overcome this limitation, our study aims to evaluate the potential of the multi-sensor synchronized RGB-D camera with depth information for maize leaf disease classification. We distinguished maize leaves from the background based on the RGB-D depth information to eliminate interference from complex field environments. Four deep learning models (i.e., Resnet50, MobilenetV2, Vgg16, and Efficientnet-B3) were used to classify three main types of maize diseases, i.e., the curvularia leaf spot [ Curvularia lunata (Wakker) Boedijn], the small spot [ Bipolaris maydis (Nishik.) Shoemaker], and the mixed spot diseases. We finally compared the pre-segmentation and post-segmentation results to test the robustness of the above models. Our main findings are: 1) The maize disease classification models based on the pre-segmentation image data performed slightly better than the ones based on the post-segmentation image data. 2) The pre-segmentation models overestimated the accuracy of disease classification due to the complexity of the background, but post-segmentation models focusing on leaf disease features provided more practical results with shorter prediction times. 3) Among the post-segmentation models, the Resnet50 and MobilenetV2 models showed similar accuracy and were better than the Vgg16 and Efficientnet-B3 models, and the MobilenetV2 model performed better than the other three models in terms of the size and the single image prediction time. Overall, this study provides a novel method for maize leaf disease classification using the post-segmentation image data from a multi-sensor synchronized RGB-D camera and offers the possibility of developing relevant portable devices.
Why it matches plant phenotyping methodsRGB-D画像による葉の病徴分類とセグメンテーション、モデル比較を中心に扱う植物病害表現型の取得・推定手法であり、単なる病理実験の測定ではない。
abstractour study aims to evaluate the potential of the multi-sensor synchronized RGB-D camera with depth information for maize leaf disease classification.
The measurement of fruit size is of great interest to estimate the yield and predict the harvest resources in advance. This work proposes a novel technique for in-field apple detection and measurement based on Deep Neural Networks. The proposed framework was trained with RGB-D data and consists of an end-to-end multitask Deep Neural Network architecture specifically designed to perform the following tasks: 1) detection and segmentation of each fruit from its surroundings; 2) estimation of the diameter of each detected fruit. The methodology was tested with a total of 15,335 annotated apples at different growth stages, with diameters varying from 27 mm to 95 mm. Fruit detection results reported an F1-score for apple detection of 0.88 and a mean absolute error of diameter estimation of 5.64 mm. These are state-of-the-art results with the additional advantages of: a) using an end-to-end multitask trainable network; b) an efficient and fast inference speed; and c) being based on RGB-D data which can be acquired with affordable depth cameras. On the contrary, the main disadvantage is the need of annotating a large amount of data with fruit masks and diameter ground truth to train the model. Finally, a fruit visibility analysis showed an improvement in the prediction when limiting the measurement to apples above 65% of visibility (mean absolute error of 5.09 mm). This suggests that future works should develop a method for automatically identifying the most visible apples and discard the prediction of highly occluded fruits.
Why it matches plant phenotyping methodsRGB-D画像と深層学習を用いて果実の検出・セグメンテーションおよび直径推定法を開発し、アノテーションデータで性能評価しているため、果実形質の取得手法が中心である。
abstractThis work proposes a novel technique for in-field apple detection and measurement based on Deep Neural Networks.
Reproduction assets foundThe authors explicitly state that the code for their multitask Mask R-CNN diameter-regression network was made publicly available together with the annotated RGB-D apple dataset (masks, diameter ground truth, spherical mask projections) at the GRAP-UdL publication page. This is a paper-specific, public, actionable codeCode · publice, which goes from
14 14 (default pooling resolution) to 28 28. After the
deconvolution, the data is flattened and fed to a linear layer
that predicts the diameter for that mask.
The developed network was implemented in the Pytorch
framework and the code has been made publicly available
jointly with the presented dataset at http://www.grap.udl.cat/en/publications/papple_rgb-d-size-dataset/.2.2.3. Network training and inference details
a) Weight initialisation: Mask ReCNN has a set of weight
initialisations pre-trained with different backbones on
ImageNet (Deng et al., 2009). In our case, the used
weights were pre-trained with a ResNet50 backbone.
However, during the course of this projecOpen asset ↗pdf-raw-page:6 lines:1-143Plant phenotyping relevance match · UnverifiedOpenAlex · checked 15 Sept 2026
The first major contribution of the paper is the proposal of using an improved DEtection Transformer network (named R2N-DETR) and Kinect-V2 camera for detecting multiple-size peaches under orchards with varied illumination and fruit occlusion. R2N-DETR model first employed Res2Net-50 to extract a fused low-high level feature map containing fine spatial features and precise semantic information of multi-size peaches from Red-Green-Blue-Depth (RGB-D) images. Second, the encoder-decoder was performed on the feature map to obtain the global context. Finally, all detected objects were detected according to each object’s global context. For the detection of 1101 RGB-D images (imaged from two orchards over three years), the R2N-DETR model achieves an average precision of 0.944 and an average detecting time of 53 ms for each image. The developed system could provide precise visual guidance for robotic picking and contribute to improving yield prediction by providing accurate fruit counting.
Why it matches plant phenotyping methodsRGB-D撮像と改良物体検出モデルを開発・評価し、モモ果実の検出とカウントという植物器官形質を抽出する方法が中心である。
abstractThe first major contribution of the paper is the proposal of using an improved DEtection Transformer network (named R2N-DETR) and Kinect-V2 camera for detecting multiple-size peaches under orchards with varied illumination and fruit occlusion.
Nondestructive measurement technology of phenotype can provide substantial phenotypic data support for applications such as seedling breeding, management, and quality testing. The current method of measuring seedling phenotypes mainly relies on manual measurement which is inefficient, subjective and destroys samples. Therefore, the paper proposes a nondestructive measurement method for the canopy phenotype of the watermelon plug seedlings based on deep learning. The Azure Kinect was used to shoot canopy color images, depth images, and RGB-D images of the watermelon plug seedlings. The Mask-RCNN network was used to classify, segment, and count the canopy leaves of the watermelon plug seedlings. To reduce the error of leaf area measurement caused by mutual occlusion of leaves, the leaves were repaired by CycleGAN, and the depth images were restored by image processing. Then, the Delaunay triangulation was adopted to measure the leaf area in the leaf point cloud. The YOLOX target detection network was used to identify the growing point position of each seedling on the plug tray. Then the depth differences between the growing point and the upper surface of the plug tray were calculated to obtain plant height. The experiment results show that the nondestructive measurement algorithm proposed in this paper achieves good measurement performance for the watermelon plug seedlings from the 1 true-leaf to 3 true-leaf stages. The average relative error of measurement is 2.33% for the number of true leaves, 4.59% for the number of cotyledons, 8.37% for the leaf area, and 3.27% for the plant height. The experiment results demonstrate that the proposed algorithm in this paper provides an effective solution for the nondestructive measurement of the canopy phenotype of the plug seedlings.
Why it matches plant phenotyping methodsスイカ苗の葉数・葉面積・草丈などの表現型を、RGB-D画像と深層学習・画像処理で非破壊測定する手法の開発が中心である。
abstractthe paper proposes a nondestructive measurement method for the canopy phenotype of the watermelon plug seedlings based on deep learning.
The accurate and rapid acquisition of fruit tree canopy parameters is fundamental for achieving precision operations in orchard robotics, including accurate spraying and precise fertilization. In response to the issue of inaccurate citrus tree canopy segmentation in complex orchard backgrounds, this paper proposes an improved DeepLabv3+ model for fruit tree canopy segmentation, facilitating canopy parameter calculation. The model takes the RGB-D (Red, Green, Blue, Depth) image segmented canopy foreground as input, introducing Dilated Spatial Convolution in Atrous Spatial Pyramid Pooling to reduce computational load and integrating Convolutional Block Attention Module and Coordinate Attention for enhanced edge feature extraction. MobileNetV3-Small is utilized as the backbone network, making the model suitable for embedded platforms. A citrus tree canopy image dataset was collected from two orchards in distinct regions. Data from Orchard A was divided into training, validation, and test set A, while data from Orchard B was designated as test set B, collectively employed for model training and testing. The model achieves a detection speed of 32.69 FPS on Jetson Xavier NX, which is six times faster than the traditional DeepLabv3+. On test set A, the mIoU is 95.62%, and on test set B, the mIoU is 92.29%, showing a 1.12% improvement over the traditional DeepLabv3+. These results demonstrate the outstanding performance of the improved DeepLabv3+ model in segmenting fruit tree canopies under different conditions, thus enabling precise spraying by orchard spraying robots.
Why it matches plant phenotyping methodsRGB-D画像から果樹キャノピーを抽出し、キャノピー形状パラメータ算出に用いる改良セグメンテーション手法を開発・検証しており、植物形質取得が中心である。
abstractThe accurate and rapid acquisition of fruit tree canopy parameters is fundamental for achieving precision operations in orchard robotics
Field / plotRGB-D / ToFLeafMorphology / geometry measurementSegmentationLeaf traits
Plant phenotyping is a research area concerned with the quantitative measurement of a plant’s structural and functional properties. In the case of measuring a leaf’s surface area, it is more often than not laborious as well as stressful to the plant. In this paper, we present the use of the RGB-D sensor, Kinect v2 as part of a portable device for non-destructive measurements of individual leaf areas (cm²/leaf) outdoors in daylight. The Kinect v2 was utilized to capture a single viewpoint 2.5D frame of plant foliage. An unsupervised clustering method, HDBSCAN was used to segment out individual leaves from the captured 2.5D frame of the subject plant. Performance of the leaf segmentation was measured by evaluating the 10 nearest (max) clusters from the sensor for each frame into 3 different categories, individual leaves (non-occluded, occluded), under-segmented and over-segmented. Probability of segmenting individual leaves differs from plant to plant, ranging from a low of 0.7178 to a high of 0.8975. The surface area of all individual non-occluded leaves obtained via the segmentation method was calculated and compared to its ground truth. The calculated individual leaf surface areas R² was recorded to range from 0.792 to 0.911 with respect to its best fit regression line while the RMSE range from 4.9482 to 14.4941 cm². The proposed system and method was shown to be capable of segmenting individual leaves from dense foliage and measuring its surface area.
Why it matches plant phenotyping methodsRGB-Dセンサーと画像分割を用いて個葉面積を非破壊測定する装置・手法を開発し、セグメンテーション性能と地上真値との精度を検証しており、植物フェノタイピング手法が中心である。
abstractIn this paper, we present the use of the RGB-D sensor, Kinect v2 as part of a portable device for non-destructive measurements of individual leaf areas (cm²/leaf) outdoors in daylight.
Improving soybean ( Glycine max L. (Merr.)) yield is crucial for strengthening national food security. Predicting soybean yield is essential to maximize the potential of crop varieties. Non-destructive methods are needed to estimate yield before crop maturity. Various approaches, including the pod-count method, have been used to predict soybean yield, but they often face issues with the crop background color. To address this challenge, we explored the application of a depth camera to real-time filtering of RGB images, aiming to enhance the performance of the pod-counting classification model. Additionally, this study aimed to compare object detection models (YOLOV7 and YOLOv7-E6E) and select the most suitable deep learning (DL) model for counting soybean pods. After identifying the best architecture, we conducted a comparative analysis of the model's performance by training the DL model with and without background removal from images. Results demonstrated that removing the background using a depth camera improved YOLOv7's pod detection performance by 10.2% precision, 16.4% recall, 13.8% mAP@50, and 17.7% mAP@0.5:0.95 score compared to when the background was present. Using a depth camera and the YOLOv7 algorithm for pod detection and counting yielded a mAP@0.5 of 93.4% and mAP@0.5:0.95 of 83.9%. These results indicated a significant improvement in the DL model's performance when the background was segmented, and a reasonably larger dataset was used to train YOLOv7.
Why it matches plant phenotyping methods深度カメラと物体検出モデルを用いてダイズ莢数という植物形質を非破壊推定する手法を開発・比較・評価しており、表現型取得が中心的である。
abstractwe explored the application of a depth camera to real-time filtering of RGB images, aiming to enhance the performance of the pod-counting classification model.
Reproduction assets foundThe paper's Data Availability Statement points to an authors' public GitHub repository containing the datasets generated and analyzed (soybean depth-camera images and pod-count segmentation data). Other URLs (labelImg, scikit-learn, CC license) are generic tools/licenses, not paper-specific assets.Dataset · publicThe datasets generated and analyzed for this study can be found in the Github repository Soybean pod count depth segmentation project 2022 accessible at https://github.com/jithin8mathew/soybean_pod_count_Depth_segmentation_project (accessed on 28 June 2023).Open asset ↗jithin8mathew/soybean_pod_count_Depth_segmentation_projectlines:183-198Plant phenotyping relevance match · UnverifiedCrossref · checked 13 Sept 2026
The precise detection and positioning of tea buds are among the major issues in tea picking automation. In this study, a novel algorithm for detecting tea buds and estimating their poses in a field environment was proposed by using a depth camera. This algorithm introduces some improvements to the YOLOv5l architecture. A Coordinate Attention Mechanism (CAM) was inserted into the neck part to accurately position the elements of interest, a BiFPN was used to enhance the small object detection ability, and a GhostConv module replaced the original Conv module in the backbone to reduce the model size and speed up model inference. After testing, the proposed detection model achieved an mAP of 85.2%, a speed of 87.71 FPS, a parameter number of 29.25 M, and a FLOPs value of 59.8 G, which are all better than those achieved with the original model. Next, an optimal pose-vertices search method (OPVSM) was developed to estimate the pose of tea by constructing a graph model to fit the pointcloud. This method could accurately estimate the poses of tea buds, with an overall accuracy of 90%, and it was more flexible and adaptive to the variations in tea buds in terms of size, color, and shape features. Additionally, the experiments demonstrated that the OPVSM could correctly establish the pose of tea buds through pointcloud downsampling by using voxel filtering with a 2 mm × 2 mm × 1 mm grid, and this process could effectively reduce the size of the pointcloud to smaller than 800 to ensure that the algorithm could be run within 0.2 s. The results demonstrate the effectiveness of the proposed algorithm for tea bud detection and pose estimation in a field setting. Furthermore, the proposed algorithm has the potential to be used in tea picking robots and also can be extended to other crops and objects, making it a valuable tool for precision agriculture and robotic applications.
Why it matches plant phenotyping methods茶芽の検出と3D姿勢推定という植物器官の形態・状態を取得する画像／深度センシング手法を開発し、精度・速度を評価しており、フェノタイピング手法が中心である。
abstracta novel algorithm for detecting tea buds and estimating their poses in a field environment was proposed by using a depth camera.
Taking maize seedlings as the object, the implementation of crops 3D reconstruction based on RGB-D binocular vision and the selection of some key parameters are investigated in this research. First, multiple images are taken from different angles around the target. By mapping the maize seedling region coordinate values after the Otsu algorithm and global threshold segmentation to the corresponding depth image, the depth data of the maize seedling region can be obtained accurately. An improved mean filter is proposed to adaptively fill the holes in the depth image. Then, the different point clouds with the fixed step angle of the maize seedling are registered and fuzed. Finally, after the fusion point cloud is simplified, the 3D model of crops can be reconstructed. Experimental results show that the simplification effect of the octree algorithm is better than that of the voxel grid filter. Among all the step angles, the reconstruction error of the step angle with 60° is the smallest. Under this condition, the height error between the model and the maize seedling is 2.22%, and the error in stem diameter is 11.67%.
Why it matches plant phenotyping methodsRGB-D双目视觉三维重建方法是论文核心，并对玉米幼苗高度和茎径等表型测量误差进行了验证。
abstractthe implementation of crops 3D reconstruction based on RGB-D binocular vision and the selection of some key parameters are investigated
Field / plotPhotogrammetry / SfM / MVSLiDAR / point cloudRGB-D / ToFStem / branchMorphology / geometry measurementSegmentation
Monitoring biodiversity in forests is crucial for their management and preservation, especially in light of increasing climatic disturbances. However, traditional methods of surveying forest biodiversity, such as the inventory of tree-related microhabitats (TreMs), are costly and time-consuming. For many years, terrestrial laser scanning (TLS) was the main method for producing highly accurate 3D models of forests. However, with recent advancements in 3D scanning technologies, there are now numerous alternatives available on the market. The aim of this study was to evaluate the performance of four different 3D data acquisition methods, i.e. close-range photogrammetry (CRP), fish-eye photogrammetry (FEP), mobile laser scanning (MLS), and mixed reality depth camera (MRDC), in terms of accuracy and ability to measure biodiversity (TreMs) at tree-stem level, in comparison to TLS. Analysis was performed based on geometric accuracy and point neighbourhood relevance. CRP was the most accurate alternative to TLS for TreM measurement with a median error of 1.5 cm, while FEP provided a good balance between accuracy (median error 1.4 cm) and speed of data collection. Although MLS showed promising results (median error 1.6 cm), noise in the point cloud limited its ability to identify TreMs. MRDC, on the other hand, had lower quality (median error 3.6 cm) and lower point density, making it unsuitable for TreM segmentation. Nevertheless, the study demonstrated the feasibility of augmenting the real world with virtual content at single-tree-stem level using mixed reality technology. Overall, the 3D scanning technologies presented hold great promise for recording the evolution of biodiversity at stem level.
Why it matches plant phenotyping methods樹幹上の生物多様性（TreMs）を対象に、複数の3D取得法をTLSと比較し、幾何精度と測定能力を評価しており、植物状態の取得手法が研究の中心である。
abstractThe aim of this study was to evaluate the performance of four different 3D data acquisition methods, i.e. close-range photogrammetry (CRP), fish-eye photogrammetry (FEP), mobile laser scanning (MLS), and mixed reality depth camera (MRDC), in terms of accuracy and ability to measure biodiversity (TreMs) at tree-stem level, in comparison to TLS.
Artificial intelligence applications enable farmers to optimize crop growth and production while reducing costs and environmental impact. Computer vision-based algorithms in particular, are commonly used for fruit segmentation, enabling in-depth analysis of the harvest quality and accurate yield estimation. In this paper, we propose TomatoDIFF, a novel diffusion-based model for semantic segmentation of on-plant tomatoes. When evaluated against other competitive methods, our model demonstrates state-of-the-art (SOTA) performance, even in challenging environments with highly occluded fruits. Additionally, we introduce Tomatopia, a new, large and challenging dataset of greenhouse tomatoes. The dataset comprises high-resolution RGB-D images and pixel-level annotations of the fruits.
Why it matches plant phenotyping methods植物上のトマト果実を画像からセグメンテーションする手法を開発・比較し、RGB-D画像と画素アノテーションのデータセットも提供しており、植物器官の状態・位置推定に関わる方法が中心である。
abstractwe propose TomatoDIFF, a novel diffusion-based model for semantic segmentation of on-plant tomatoes
Reproduction assets foundThe paper introduces TomatoDIFF and the Tomatopia dataset, with explicit public availability of source code and dataset at the authors' GitHub repository. It also trains/evaluates on the public Kaggle 'Tomato dataset' (andrewmvd/tomato-detection), which is a paper-specific public image dataset used directly in the phenCode · publicThe source code of TomatoDIFF and Tomatopia are available at https://github.com/MIvanovska/TomatoDIFF .Open asset ↗MIvanovska/TomatoDIFFlines:1-44Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Published1 Jul 2023Computers and Electronics in Agriculture.
Several variables, including a rising human population, varying weather patterns in the context of ongoing climate change, and the rapid worldwide spread of epidemics, all contribute to boosting agricultural demand. To assure food availability, quality, and safety while increasing yields and profitability, precision agriculture must progress swiftly. Precision viticulture aims to optimize vineyard management in this setting by reducing resource consumption and environmental impact while simultaneously enhancing the yield, product quality, and oenological potential of vineyards. This comprehensive review article offers an overview of the real-world and laboratory applications of optical and non-optical sensors in precision viticulture for 3D modelling. Hence, there is a pressing need to track the development of crops at a wide range of spatial and temporal scales, in a wide variety of environments, and for a wide range of objectives in a non-destructive manner. Due to the intrinsic spatial heterogeneity of vineyards, the adoption of precision viticulture necessitates crop monitoring using contactless and non-invasive sensors such as ultrasonic, LiDAR (Light Detection and Ranging), depth, or RGB cameras to prevent low accuracy and sparse sampling. This study aims to assist researchers in gaining a broad understanding of the sensing technologies for precision viticulture, the present problems, and the advancement of the state of the art. The study focuses on sensors used for Proximal Sensing to geometrically characterize vines using statically or dynamically ground-based measurements through a wide range of mobile sensing platforms. The employed sensors, data extraction, and analysis procedures are described. Moreover, the present and future potential of Proximal Sensing and Remote Sensing in vineyards is discussed.
Why it matches plant phenotyping methodsブドウ樹の形状を光学・非光学センサーで計測・3Dモデル化する近接センシング手法を中心に扱うレビューであり、植物フェノタイピング手法レビューに該当する。
abstractThis comprehensive review article offers an overview of the real-world and laboratory applications of optical and non-optical sensors in precision viticulture for 3D modelling.
In this work, a Multispectral-Depth (MS-D) imaging system for close range plant inspection is presented. The proposed system is comprised of multispectral cameras calibrated with respect to an RGB-D camera. It is ultimately tested and quantitatively compared to state-of-the-art feature-based methods for multispectral image registration. The results show that MS-D system outperforms state-of-the-art methods in all experimental trials, from registration of checkerboard images (where the accuracy of the MS-D system was on a sub-pixel level) to registration of feature-rich plant images, both real and synthetic. While the greatest registration error of the MS-D system amounted to 9 pixels, registration error of the feature-based method was up to 19 times greater. Additionally, contrast to the state-of-the-art feature matching approaches, the MS-D system, once calibrated, is applicable as is, without the need for recalibration. As a part of this work, MS-D has been deployed in encapsulated growth chambers for rapid data collection and in a small indoor organic farm for automated plant monitoring. As a part of the experimental plant monitoring study, it was shown that vegetation indices calculated with the MS-D system can be used to estimate water stress in Spathiphyllum plants equally well as with the spectroradiometer, human operated device for measurement of plant vegetation indices. The biggest relative change between the vegetation indices calculated for the plants exposed to the short term water stress and the control group was found in values of NDRE, amounting to 94.8%, followed by the relative change of 64.9% observed for the values of SR.
Why it matches plant phenotyping methodsMS-Dマルチスペクトル・深度画像システムを開発し、画像レジストレーション性能を定量比較するとともに、植物モニタリングと水ストレス推定へ応用しており、植物表現型取得法が中心的である。
abstracta Multispectral-Depth (MS-D) imaging system for close range plant inspection is presented
An operator of a wild blueberry harvester faces the fatigue of manually adjusting the height of the harvester's head, considering spatial variations in plant height, fruit zone, and field topography affecting fruit yield. For stress-free harvesting of wild blueberries, a deep learning-supported machine vision control system has been developed to detect the fruit height and precisely auto-adjust the header picking teeth rake position. The OpenCV AI Kit (OAK-D) was used with YOLOv4-tiny deep learning model with code developed in Python to solve the challenge of matching fruit heights with the harvester's head position. The system accuracy was statistically evaluated with R 2 (coefficient of determination) and σ (standard deviation) measured on the difference in distances between the berries picking teeth and average fruit heights, which were 72, 43% and 2.1, 2.3 cm for the auto and manual head adjustment systems, respectively. This innovative system performed well in weed-free areas but requires further work to operate in weedy sections of the fields. Benefits of using this system include automated control of the harvester's head to match the header picking rake height to the level of the fruit height while reducing the operator's stress by creating safer working environments.
Why it matches plant phenotyping methods果実の高さという植物器官の形態形質を画像から推定し、収穫機ヘッダーを自動調整する機械視覚システムを開発・評価しており、単なる収穫対象の位置検出を超えた形質取得が中心である。
abstracta deep learning-supported machine vision control system has been developed to detect the fruit height and precisely auto-adjust the header picking teeth rake position.
Field / plotRGB-D / ToFStem / branchObject detection2D/3D reconstruction
Abstract In this work, motivated by the precision agriculture (PA) paradigm, we address the problem of managing hazelnut suckering plants on a per‐plant basis in a large‐scale orchard. Suckering plants, or shortly, suckers, are basal shoots that grow at the base of a tree and compete with the tree itself for nutrients and water. Generally, in large‐scale orchards, suckers are treated with the application of herbicide through spraying tractors that continuously spray the crops while navigating the whole orchard. This approach however does not consider the individual needs of each plant and it is definitely not environmentally‐friendly since a lot of unnecessary solution is being drained in the soil. For this reason, we propose a novel fully autonomous sucker management architecture that is able to detect the presence of suckers for each plant, by relying on a You Only Look Once (YOLO)‐based recognition system, reconstruct them in three‐dimension and estimate the amount of herbicide solution needed for the specific plant, based on a data‐driven approach. The herbicide solution is applied using a ground robot equipped with an RGB‐D camera and a spraying system. This approach allows to significantly reduce pollution and waste. Experimental results both for individual components and for the entire architecture in a real‐world (1:1 scale) hazelnut orchard located in Caprarola, Italy, are provided to corroborate the proposed architecture.
Why it matches plant phenotyping methods個体ごとの吸枝の検出と3次元再構成により植物状態を推定する画像ベースの自律センシング・散布アーキテクチャが研究の中心であり、単なる農業管理への応用を超えている。
abstractwe propose a novel fully autonomous sucker management architecture that is able to detect the presence of suckers for each plant, by relying on a You Only Look Once (YOLO)‐based recognition system, reconstruct them in three‐dimension and estimate the amount of herbicide solution needed for the specific plant
Phenotyping plays a significant role in the breeding of apple tree. However, existing researches mainly relied on instruments, such as LiDAR, RGB-D camera or UAV (unmanned aerial vehicle) embedded with depth sensor, etc., which requires additional costs for users and also inconvenient. Therefore, a novel method of smartphone-based heterogeneous binocular vision was developed to fulfill low-cost automated phenotyping for apple tree. In this study, a pair of cameras on multi-camera smartphone was selected to obtain heterogeneous binocular camera. After that, a so-called virtual focal method was developed to generate standard binocular images from heterogeneous binocular images of smartphone. A well-known YOLOv5s object detection model was trained on a four-class dataset to detect fruits, grafts, trunks and whole trees. Then, the model was simplified to fit the deployment on smartphone. Finally, five phenotypes (trunk diameter, ground diameter, tree height, fruit vertical diameter, and fruit horizontal diameter) of individual apple tree were obtained by pinhole camera model and standard binocular vision. After evaluation of phenotyping manually and by smartphone, our method shows MAPE (mean average percentage error) ranging from 6.00 % to 13.73 % for the five phenotypes. Compared with the existing studies, our method has reached a close or even better phenotyping accuracy with only a smartphone. As more and more smartphones have multi-camera, our method is probably the lowest cost phenotyping method for most of the potential users. Results indicated that the approach could be utilized to phenotyping of apple tree.
Why it matches plant phenotyping methodsスマートフォンのステレオ画像とYOLOv5sを用いて、リンゴ樹の複数形質を自動推定する手法を開発・評価しており、表現型取得が研究の中心である。
abstracta novel method of smartphone-based heterogeneous binocular vision was developed to fulfill low-cost automated phenotyping for apple tree.
Water plays a very important role in the growth of tomato ( Solanum lycopersicum L.), and how to detect the water status of tomato is the key to precise irrigation. The objective of this study is to detect the water status of tomato by fusing RGB, NIR and depth image information through deep learning. Five irrigation levels were set to cultivate tomatoes in different water states, with irrigation amounts of 150%, 125%, 100%, 75%, and 50% of reference evapotranspiration calculated by a modified Penman-Monteith equation, respectively. The water status of tomatoes was divided into five categories: severely irrigated deficit, slightly irrigated deficit, moderately irrigated, slightly over-irrigated, and severely over-irrigated. RGB images, depth images and NIR images of the upper part of the tomato plant were taken as data sets. The data sets were used to train and test the tomato water status detection models built with single-mode and multimodal deep learning networks, respectively. In the single-mode deep learning network, two CNNs, VGG-16 and Resnet-50, were trained on a single RGB image, a depth image, or a NIR image for a total of six cases. In the multimodal deep learning network, two or more of the RGB images, depth images and NIR images were trained with VGG-16 or Resnet-50, respectively, for a total of 20 combinations. Results showed that the accuracy of tomato water status detection based on single-mode deep learning ranged from 88.97% to 93.09%, while the accuracy of tomato water status detection based on multimodal deep learning ranged from 93.09% to 99.18%. The multimodal deep learning significantly outperformed the single-modal deep learning. The tomato water status detection model built using a multimodal deep learning network with ResNet-50 for RGB images and VGG-16 for depth and NIR images was optimal. This study provides a novel method for non-destructive detection of water status of tomato and gives a reference for precise irrigation management.
Why it matches plant phenotyping methodsRGB・NIR・深度画像を融合した深層学習により、トマトの水分状態という植物生理状態を非破壊推定する方法が研究の中心であり、モデル比較と精度評価も行っている。
abstractThe objective of this study is to detect the water status of tomato by fusing RGB, NIR and depth image information through deep learning.
Field / plotRGB-D / ToFStem / branch2D/3D reconstructionSegmentationSkeletonization / topology
The shortage of agricultural labourers worldwide has left many groups unharvested and wasted, which motivates researchers worldwide to research fruit harvesting robots extensively. One of the major problems in fruit harvesting is to selectively avoid hard obstacles such as tree branches so that more optimal picking positions can be found and more fruits throughout the tree can be harvested. However, tree branches are often obscured in unstructured natural orchards and thus necessary branch reconstruction and recovery are required. The current branch reconstruction and recovery methods for harvesting robots focus on planar reconstruction with few occlusions while the existing 3D tree modelling methods are not optimised for harvesting purposes that require low computational cost and high localisation accuracy. This work presented a novel framework that reconstructs and recovers 3D obscured branches from planar images and depth maps captured by an RGB-D camera. The framework comprises three parts: branch segmentation using Unet++, branch reconstruction using Point2Skeleton and branch recovery using a novel obscured branch recovery (OBR) algorithm. Branch segmentation using Unet++ with InceptionV3 encoder shows the best overall result with IoU and F1-score of 0.6249 and 0.7692 respectively. OBR recovery algorithm achieves average reconstruction accuracy of 0.72. The mean error of the reconstructed total surface and obscured surface using OBR is 18.68 mm and 38.11 mm with a standard deviation of 14.3 mm and 32.64 mm. The result shows that this framework can effectively reconstruct spatial information of visible and obscured branches from a single view image which can potentially be utilised in harvesting robots.
Why it matches plant phenotyping methodsRGB-D画像から枝の分割・3D再構成・遮蔽部復元を行う技術が中心で、単なる収穫対象の位置検出を超えて植物器官の空間構造を定量化しているため、植物表現型計測法として含める。
abstractThis work presented a novel framework that reconstructs and recovers 3D obscured branches from planar images and depth maps captured by an RGB-D camera.
Labor shortages in fruit crop production have prompted the development of mechanized and automated machines as alternatives to labor-intensive orchard operations such as harvesting, pruning, and thinning. Agricultural robots capable of identifying tree canopy parts and estimating geometric and topological parameters, such as branch diameter, length, and angles, can optimize crop yields through automated pruning and thinning platforms. In this study, we proposed a machine vision system to estimate canopy parameters in apple orchards and determine an optimal number of fruit for individual branches, providing a foundation for robotic pruning, flower thinning, and fruitlet thinning to achieve desired yield and quality.Using color and depth information from an RGB-D sensor (Microsoft Azure Kinect DK), a YOLOv8-based instance segmentation technique was developed to identify trunks and branches of apple trees during the dormant season. Principal Component Analysis was applied to estimate branch diameter (used to calculate limb cross-sectional area, or LCSA) and orientation. The estimated branch diameter was utilized to calculate LCSA, which served as an input for crop-load estimation, with larger LCSA values indicating a higher potential fruit-bearing capacity.RMSE for branch diameter estimation was 2.08 mm, and for crop-load estimation, 3.95. Based on commercial apple orchard management practices, the target crop-load (number of fruit) for each segmented branch was estimated with a mean absolute error (MAE) of 2.99 (ground truth crop-load was 6 apples per LCSA). This study demonstrated a promising workflow with high performance in identifying trunks and branches of apple trees in dynamic commercial orchard environments and integrating farm management practices into automated decision-making.
Why it matches plant phenotyping methodsRGB-D画像とYOLOv8を用いてリンゴ樹の枝形態（直径・方向）を抽出し、樹体の作物負荷を推定する手法を開発・評価しており、表現型取得が研究の中心である。
abstractIn this study, we proposed a machine vision system to estimate canopy parameters in apple orchards and determine an optimal number of fruit for individual branches
The detection and sizing of fruits with computer vision methods is of interest because it provides relevant information to improve the management of orchard farming. However, the presence of partially occluded fruits limits the performance of existing methods, making reliable fruit sizing a challenging task. While previous fruit segmentation works limit segmentation to the visible region of fruits (known as modal segmentation), in this work we propose an amodal segmentation algorithm to predict the complete shape, which includes its visible and occluded regions. To do so, an end-to-end convolutional neural network (CNN) for simultaneous modal and amodal instance segmentation was implemented. The predicted amodal masks were used to estimate the fruit diameters in pixels. Modal masks were used to identify the visible region and measure the distance between the apples and the camera using the depth image. Finally, the fruit diameters in millimetres (mm) were computed by applying the pinhole camera model. The method was developed with a Fuji apple dataset consisting of 3925 RGB-D images acquired at different growth stages with a total of 15,335 annotated apples, and was subsequently tested in a case study to measure the diameter of Elstar apples at different growth stages. Fruit detection results showed an F1-score of 0.86 and the fruit diameter results reported a mean absolute error (MAE) of 4.5 mm and R2 = 0.80 irrespective of fruit visibility. Besides the diameter estimation, modal and amodal masks were used to automatically determine the percentage of visibility of measured apples. This feature was used as a confidence value, improving the diameter estimation to MAE = 2.93 mm and R2 = 0.91 when limiting the size estimation to fruits detected with a visibility higher than 60%. The main advantages of the present methodology are its robustness for measuring partially occluded fruits and the capability to determine the visibility percentage. The main limitation is that depth images were generated by means of photogrammetry methods, which limits the efficiency of data acquisition. To overcome this limitation, future works should consider the use of commercial RGB-D sensors. The code and the dataset used to evaluate the method have been made publicly available at https://github.com/GRAP-UdL-AT/Amodal_Fruit_Sizing.
Why it matches plant phenotyping methods果実の遮蔽に頑健な画像ベースのアモーダル分割と、リンゴ果径という植物形質の推定手法を開発・検証しており、方法が研究の中心である。
abstractThe predicted amodal masks were used to estimate the fruit diameters in pixels.
Reproduction assets foundThe paper's apple amodal segmentation dataset (RGB-D images, modal/amodal masks, calliper-measured diameters) and the authors' analysis code are both explicitly stated to be publicly available at the authors' GitHub repository GRAP-UdL-AT/Amodal_Fruit_Sizing.Dataset · publictain data from both maturity stages, of different fruit size and with
different fruit visibilities. The dataset split was performed randomly,
obtaining in each partition a similar distribution of diameters (Fig. 4.b)
and apples visibilities (Fig. 4.d) than in the original dataset. The dataset
has been made publicly available at https://github.com/GRAP-UdL-AT/Amodal_Fruit_Sizing.The data used for the case study was acquired in an Elstar apple
orchard located in Randwijk (the Netherlands). Five different trees were
imaged at four different dates (Table 1), obtaining data at different
growth stages: BBCH75, BBCH77, BBCH78 and BBCH85 (Fig. 2b). To
have a complete representation of trees, images Open asset ↗GRAP-UdL-AT/Amodal_Fruit_Sizingpdf-raw-page:3 lines:1-74Code · publicft, Supervision.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Data availability
The code and the dataset used to evaluate the method have been
made publicly available at https://github.com/GRAP-UdL-AT/Amodal_Fruit_Sizing.Acknowledgements
This work was partly funded by the Departament de Recerca i Uni
versitats de la Generalitat de Catalunya (grant 2021 LLAV 00088), the
Spanish Ministry of Science, Innovation and Universities (grants
RTI2018-094222-B-I00 [PAgFRUIT project], PID2021-126648OB-I00
[PAgPROTECT project] and PID2020-117142GOpen asset ↗GRAP-UdL-AT/Amodal_Fruit_Sizingpdf-raw-page:12 lines:1-75Plant phenotyping relevance match · UnverifiedOpenAlex · checked 13 Sept 2026
Three colour and depth (RGB-D) devices were compared, to assess the effect of depth image misalignment, resulting from simultaneous localisation and mapping (SLAM) error, due to forest structure complexity. Urban parkland (S1) was used to assess stem density, and understory vegetation (≤1.3 m) was assessed in native woodland (S2). Individual stem and continuous capture approaches were used, with stem diameter at breast height (DBH) estimated. Misalignment was present within point clouds; however, no significant differences in DBH were observed for stems captured at S1 with either approach (Kinect p = 0.16; iPad p = 0.27; Zed p = 0.79). Using continuous capture, the iPad was the only RGB-D device to maintain SLAM in all S2 plots. There was significant correlation between DBH error and surrounding understory vegetation with the Kinect device (p = 0.04). Conversely, there was no significant relationship between DBH error and understory vegetation for the iPad (p = 0.55) and Zed (p = 0.86). The iPad had the lowest DBH root-mean-square error (RMSE) across both individual stem (RMSE = 2.16cm) and continuous (RMSE = 3.23cm) capture approaches. The results suggest that the assessed RGB-D devices are more capable of operation within complex forest environments than previous generations.
Why it matches plant phenotyping methodsRGB-D深度センサーを比較・検証し、森林環境での樹木DBHという植物形態形質の推定精度とSLAM性能を評価しており、フェノタイピング手法が中心です。
abstractThree colour and depth (RGB-D) devices were compared, to assess the effect of depth image misalignment
Accurate and high-throughput plant phenotyping is important for accelerating crop breeding. Spectral imaging that can acquire both spectral and spatial information of plants related to structural, biochemical, and physiological traits becomes one of the popular phenotyping techniques. However, close-range spectral imaging of plants could be highly affected by the complex plant structure and illumination conditions, which becomes one of the main challenges for close-range plant phenotyping. In this study, we proposed a new method for generating high-quality plant 3-dimensional multispectral point clouds. Speeded-Up Robust Features and Demons was used for fusing depth and snapshot spectral images acquired at close range. A reflectance correction method for plant spectral images based on hemisphere references combined with artificial neural network was developed for eliminating the illumination effects. The proposed Speeded-Up Robust Features and Demons achieved an average structural similarity index measure of 0.931, outperforming the classic approaches with an average structural similarity index measure of 0.889 in RGB and snapshot spectral image registration. The distribution of digital number values of the references at different positions and orientations was simulated using artificial neural network with the determination coefficient ( R 2 ) of 0.962 and root mean squared error of 0.036. Compared with the ground truth measured by ASD spectrometer, the average root mean squared error of the reflectance spectra before and after reflectance correction at different leaf positions decreased by 78.0%. For the same leaf position, the average Euclidean distances between the multiview reflectance spectra decreased by 60.7%. Our results indicate that the proposed method achieves a good performance in generating plant 3-dimensional multispectral point clouds, which is promising for close-range plant phenotyping.
Why it matches plant phenotyping methods植物の3次元マルチスペクトル点群生成と反射補正・画像融合手法を開発し、既存手法や分光計を用いて性能検証しており、植物表現型取得が研究の中心である。
abstractIn this study, we proposed a new method for generating high-quality plant 3-dimensional multispectral point clouds.
Traditionally, vegetable and fruit production has relied on empirical and ambiguous decisions made by human farmers. To overcome this uncertainty in agriculture, smart farm robots have been widely studied in recent years. However, measuring growth information with robots remains a challenge because of the similarity in the appearance of the target plant and those around it. In this study, we propose a smart farm robot that accurately measures the growth information of a target plant based on object detection, image fusion, and data augmentation with fused images. The proposed smart farm robot uses an end-to-end real-time deep learning-based object detector that shows state-of-the-art performances. To distinguish the target plant from other plants with a higher accuracy and improved robustness than those of existing methods, we exploited image fusion using both RGB and depth images. In particular, the data augmentation, based on the fused RGB, and depth information, contributes to the precise measurement of growth information from smart farms, regardless of the high density of vegetables and fruits in these farms. We propose and evaluate a real-time measurement system to obtain precise target-plant growth information in precision agriculture. The code and models are publicly available on Github: https://github.com/kistvision/Plant_growth_measurement.
Why it matches plant phenotyping methodsRGB・深度画像と物体検出を用いて対象植物の生育情報を取得するリアルタイム測定システムが研究の中心であり、植物フェノタイピング手法に該当する。
abstractwe propose a smart farm robot that accurately measures the growth information of a target plant based on object detection, image fusion, and data augmentation with fused images.
In most of the countries, grapes are considered as a cash crop. Currently huge research is going on in development of automated grape harvesting systems. Speedy and reliable grape bunch detection is prime need for various deep learning based automated systems which deals with object detection and object segmentation tasks. But currently very few datasets are available on grape bunches in vineyard, because of which there is restriction to the research in this area. In comparison to the vineyard in outside countries, Indian vineyard structure is more complex, so it becomes hard to work in real-time. To overcome these problems and to make vineyard dataset for suitable for Indian vineyard scenarios, this paper proposed four different datasets on grape bunches in vineyard. For creating all datasets in GrapesNet, natural environmental conditions have been considered. GrapesNet includes total 11000+ images of grape bunches. Necessary data for weight prediction of grape cluster is also provided with dataset like height, width and real weight of cluster present in image. Proposed datasets can be used for prime tasks like grape bunch detection, grape bunch segmentation, and grape bunch weight estimation etc. of future generation automated vineyard harvesting technologies.
Why it matches plant phenotyping methodsブドウ果房画像データセットを構築し、果房の検出・セグメンテーションに加えて重量推定用の寸法と実重量を提供することが中心で、再利用可能な植物表現型データセットに該当する。
abstractthis paper proposed four different datasets on grape bunches in vineyard.
Reproduction assets foundThe paper is a data descriptor for GrapesNet, a public Mendeley Data repository of Indian vineyard RGB/RGB-D grape bunch image datasets with ground-truth cluster height, width, and weight measurements used for phenotyping tasks (detection, segmentation, weight estimation). The dataset is the paper's core asset and is aDataset · publicRepository name: GrapesNet: Indian Grape Clusters RGB & RGB-D Image Datasets
Data identification number (DOI): 10.17632/mhzmzd5cwx.1
Direct URL to data: https://data.mendeley.com/datasets/mhzmzd5cwx/1Open asset ↗10.17632/mhzmzd5cwx.1lines:1-95Plant phenotyping relevance match · UnverifiedOpenAlex · Crossref · checked 13 Sept 2026
Solving the problem of the stem contour extraction of maize is difficult under open field conditions, and the stem diameter cannot be measured quickly and nondestructively. In this paper, maize at the small and large bell stages was the object of study. An adaptive threshold segmentation algorithm based on the color space model was proposed to obtain the stem contour and stem diameter of maize in the field. Firstly, 2D images of the maize stem in the field were captured with an RGB-D camera. Then, the images were processed by hue saturation value (HSV) color space. Next, the stem contour of the maize was extracted by maximum between-class variance (Otsu). Finally, the reference method was used to obtain the stem diameter of the maize. Scatter plots and Dice coefficients were used to compare the contour extraction effects of the HSV + fixed threshold algorithm, the HSV + Otsu algorithm, and the HSV + K-means algorithm. The results showed that the HSV + Otsu algorithm is the optimal choice for extracting the maize stem contour. The mean absolute error, mean absolute percentage error (MAPE), and root mean square error (RMSE) of the maize stem diameter at the small bell stage were 4.30 mm, 10.76%, and 5.29 mm, respectively. The mean absolute error, MAPE, and RMSE of the stem diameter of the maize at the large bell stage were 4.78 mm, 12.82%, and 5.48 mm, respectively. The MAPE was within 10–20%. The results showed that the HSV + Otsu algorithm could meet the requirements for stem diameter measurement and provide a reference for the acquisition of maize phenotypic parameters in the field. In the meantime, the acquisition of maize phenotypic parameters under open field conditions provides technical and data support for precision farming and plant breeding.
Why it matches plant phenotyping methods圃場画像からトウモロコシ茎の輪郭と直径を抽出・測定する適応的画像解析手法を開発し、複数アルゴリズムと基準法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ある。
abstractScatter plots and Dice coefficients were used to compare the contour extraction effects of the HSV + fixed threshold algorithm, the HSV + Otsu algorithm, and the HSV + K-means algorithm.
Recent studies indicate that food demand will increase by 35-56% over the period 2010-2050 due to population increase, economic development, and urbanization. Greenhouse systems allow for the sustainable intensification of food production with demonstrated high crop production per cultivation area. Breakthroughs in resource-efficient fresh food production merging horticultural and AI expertise take place with the international competition "Autonomous Greenhouse Challenge". This paper describes and analyzes the results of the third edition of this competition. The competition's goal is the realization of the highest net profit in fully autonomous lettuce production. Two cultivation cycles were conducted in six high-tech greenhouse compartments with operational greenhouse decision-making realized at a distance and individually by algorithms of international participating teams. Algorithms were developed based on time series sensor data of the greenhouse climate and crop images. High crop yield and quality, short growing cycles, and low use of resources such as energy for heating, electricity for artificial light, and CO 2 were decisive in realizing the competition's goal. The results highlight the importance of plant spacing and the moment of harvest decisions in promoting high crop growth rates while optimizing greenhouse occupation and resource use. In this paper, images taken with depth cameras (RealSense) for each greenhouse were used by computer vision algorithms (Deepabv3+ implemented in detectron2 v0.6) in deciding optimum plant spacing and the moment of harvest. The resulting plant height and coverage could be accurately estimated with an R 2 of 0.976, and a mIoU of 98.2, respectively. These two traits were used to develop a light loss and harvest indicator to support remote decision-making. The light loss indicator could be used as a decision tool for timely spacing. Several traits were combined for the harvest indicator, ultimately resulting in a fresh weight estimation with a mean absolute error of 22 g. The proposed non-invasively estimated indicators presented in this article are promising traits to be used towards full autonomation of a dynamic commercial lettuce growing environment. Computer vision algorithms act as a catalyst in remote and non-invasive sensing of crop parameters, decisive for automated, objective, standardized, and data-driven decision making. However, spectral indexes describing lettuces growth and larger datasets than the currently accessible are crucial to address existing shortcomings between academic and industrial production systems that have been encountered in this work.
Why it matches plant phenotyping methods深度カメラ画像とコンピュータビジョンによりレタスの草丈・被覆率・収量関連形質を推定し、精度評価と自動意思決定指標への応用を行っており、植物表現型取得法が中心である。
abstractimages taken with depth cameras (RealSense) for each greenhouse were used by computer vision algorithms (Deepabv3+ implemented in detectron2 v0.6) in deciding optimum plant spacing and the moment of harvest.
Reproduction assets foundThe paper's complete challenge dataset (climate time-series and annotated lettuce crop images used for the computer vision phenotyping) is published open access on 4TU.ResearchData, cited both in the Data Availability Statement and in reference 56.Dataset · public3rd Autonomous Greenhouse Challenge-Real Challenge Data Climate and Images Dataset: 4TU.ResearchData 2023 Available online: https://data.4tu.nl/articles/dataset/3rd_Autonomous_Greenhouse_Challenge_Online_Challenge_Lettuce_Images/15023088Open asset ↗4TU.ResearchData · 15023088lines:853-968Code / dataset availability confirmedEurope PMC · OpenAlex · checked 15 Sept 2026
Smart farming (SF) applications rely on robust and accurate computer vision systems. An important computer vision task in agriculture is semantic segmentation, which aims to classify each pixel of an image and can be used for selective weed removal. State-of-the-art implementations use convolutional neural networks (CNN) that are trained on large image datasets. In agriculture, publicly available RGB image datasets are scarce and often lack detailed ground-truth information. In contrast to agriculture, other research areas feature RGB-D datasets that combine color (RGB) with additional distance (D) information. Such results show that including distance as an additional modality can improve model performance further. Therefore, we introduce WE3DS as the first RGB-D image dataset for multi-class plant species semantic segmentation in crop farming. It contains 2568 RGB-D images (color image and distance map) and corresponding hand-annotated ground-truth masks. Images were taken under natural light conditions using an RGB-D sensor consisting of two RGB cameras in a stereo setup. Further, we provide a benchmark for RGB-D semantic segmentation on the WE3DS dataset and compare it with a solely RGB-based model. Our trained models achieve up to 70.7% mean Intersection over Union (mIoU) for discriminating between soil, seven crop species, and ten weed species. Finally, our work confirms the finding that additional distance information improves segmentation quality.
Why it matches plant phenotyping methods植物種の画素単位セグメンテーション用RGB-Dデータセットとベンチマークを構築し、植物識別・分離という表現型取得ワークフローを中心的に評価しているため。
abstractwe introduce WE3DS as the first RGB-D image dataset for multi-class plant species semantic segmentation in crop farming.
Reproduction assets foundThe paper's WE3DS RGB-D image dataset (2568 annotated images) and the authors' modified ESANet analysis code are publicly deposited on Zenodo (DOI 10.5281/zenodo.7457983), as stated in the experiments section. The MDPI supplementary file contains only tables (species list, depth accuracy, confusion matrices), not the影像Dataset · public024 × 512
20.6
27.0
37.7 †
34.2
11.5
RGB
1024 × 512
52.4
22.2
39.2 †
38.4
11.5
RGB-D
1024 × 512
59.1
19.2
85.8
55.3
18.5
D
1280 × 960
48.5
11.3
154.1
37.0
27.1
RGB
1280 × 960
70.1
11.0
156.1
46.8
27.1
RGB-D
1280 × 960
70.7
8.6
240.3
66.6
43.4
Information on the dataset and modified code of the ESANet can be found on our website https://doi.org/10.5281/zenodo.7457983 (accessed on 18 December 2022).
4.4. ResultsOpen asset ↗Zenodo · 10.5281/zenodo.7457983lines:69-146Plant phenotyping relevance match · UnverifiedEurope PMC · Crossref · OpenAlex · checked 7 Sept 2026
Background Plant shape and structure are important factors in peanut breeding research. Constructing a three-dimension (3D) model can provide an effective digital tool for comprehensive and quantitative analysis of peanut plant structure. Fast and accurate are always the goals of the plant 3D model reconstruction research. Results We proposed a 3D reconstruction method based on dual RGB-D cameras for the peanut plant 3D model quickly and accurately. The two Kinect v2 were mirror symmetry placed on both sides of the peanut plant, and the point cloud data obtained were filtered twice to remove noise interference. After rotation and translation based on the corresponding geometric relationship, the point cloud acquired by the two Kinect v2 was converted to the same coordinate system and spliced into the 3D structure of the peanut plant. The experiment was conducted at various growth stages based on twenty potted peanuts. The plant traits' height, width, length, and volume were calculated through the reconstructed 3D models, and manual measurement was also carried out during the experiment processing. The accuracy of the 3D model was evaluated through a synthetic coefficient, which was generated by calculating the average accuracy of the four traits. The test result showed that the average accuracy of the reconstructed peanut plant 3D model by this method is 93.42%. A comparative experiment with the iterative closest point (ICP) algorithm, a widely used 3D modeling algorithm, was additionally implemented to test the rapidity of this method. The test result shows that the proposed method is 2.54 times faster with approximated accuracy compared to the ICP method. Conclusions The reconstruction method for the 3D model of the peanut plant described in this paper is capable of rapidly and accurately establishing a 3D model of the peanut plant while also meeting the modeling requirements for other species' breeding processes. This study offers a potential tool to further explore the 3D model for improving traits and agronomic qualities of plants.
Why it matches plant phenotyping methodsピーナッツ植物の3D再構成手法を開発し、植物形質の定量と手動測定・ICP法による精度および速度比較で検証しており、フェノタイピング手法が研究の中心である。
abstractWe proposed a 3D reconstruction method based on dual RGB-D cameras for the peanut plant 3D model quickly and accurately.
Abstract Current techniques of forest inventory rely on manual measurements and are slow and labor intensive. Recent developments in computer vision and depth sensing can produce accurate measurement data at significantly reduced time and labor costs. We developed the ForSense system to measure the diameters of trees at various points along the stem as well as stem straightness. Time use, mean absolute error (MAE), and root mean squared error (RMSE) metrics were used to compare the system against manual methods, and to compare the system against itself (reproducibility). Depth-derived diameter measurements of the stems at the heights of 0.3, 1.4, and 2.7 m achieved RMSE of 1.7, 1.5, and 2.7 cm, respectively. The ForSense system produced straightness measurement data that was highly correlated with straightness ratings by trained foresters. The ForSense system was also consistent, achieving sub-centimeter diameter difference with subsequent measures and less than 4% difference in straightness value between runs. This method of forest inventory, which is based on depth-image computer vision, is time efficient compared to manual methods and less computationally and technologically intensive compared to Structure-from-Motion (SFM) photogrammetry and ground-based LiDAR or terrestrial laser scanning (TLS).
Why it matches plant phenotyping methods深度画像コンピュータビジョンで樹木の幹径と幹の通直性を測定するForSenseシステムを開発し、手動測定との精度比較と再現性評価を行っており、植物形質の取得手法が研究の中心である。
abstractWe developed the ForSense system to measure the diameters of trees at various points along the stem as well as stem straightness.
Stereo matching is a depth perception method for plant phenotyping with high throughput. In recent years, the accuracy and real-time performance of the stereo matching models have been greatly improved. While the training process relies on specialized large-scale datasets, in this research, we aim to address the issue in building stereo matching datasets. A semi-automatic method was proposed to acquire the ground truth, including camera calibration, image registration, and disparity image generation. On the basis of this method, spinach, tomato, pepper, and pumpkin were considered for experiment, and a dataset named PlantStereo was built for reconstruction. Taking data size, disparity accuracy, disparity density, and data type into consideration, PlantStereo outperforms other representative stereo matching datasets. Experimental results showed that, compared with the disparity accuracy at pixel level, the disparity accuracy at sub-pixel level can remarkably improve the matching accuracy. More specifically, for PSMNet, the EPE and bad−3 error decreased 0.30 pixels and 2.13%, respectively. For GwcNet, the EPE and bad−3 error decreased 0.08 pixels and 0.42%, respectively. In addition, the proposed workflow based on stereo matching can achieve competitive results compared with other depth perception methods, such as Time-of-Flight (ToF) and structured light, when considering depth error (2.5 mm at 0.7 m), real-time performance (50 fps at 1046 × 606), and cost. The proposed method can be adopted to build stereo matching datasets, and the workflow can be used for depth perception in plant phenotyping.
Why it matches plant phenotyping methods植物フェノタイピング用のステレオマッチングデータセット構築法を開発し、精度・性能を検証した研究であり、表現型取得手法が中心である。
abstractStereo matching is a depth perception method for plant phenotyping with high throughput.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · Crossref · checked 7 Sept 2026
Introduction: Nondestructive detection of crop phenotypic traits in the field is very important for crop breeding. Ground-based mobile platforms equipped with sensors can efficiently and accurately obtain crop phenotypic traits. In this study, we propose a dynamic 3D data acquisition method in the field suitable for various crops by using a consumer-grade RGB-D camera installed on a ground-based movable platform, which can collect RGB images as well as depth images of crop canopy sequences dynamically. Methods: A scale-invariant feature transform (SIFT) operator was used to detect adjacent date frames acquired by the RGB-D camera to calculate the point cloud alignment coarse matching matrix and the displacement distance of adjacent images. The data frames used for point cloud matching were selected according to the calculated displacement distance. Then, the colored ICP (iterative closest point) algorithm was used to determine the fine matching matrix and generate point clouds of the crop row. The clustering method was applied to segment the point cloud of each plant from the crop row point cloud, and 3D phenotypic traits, including plant height, leaf area and projected area of individual plants, were measured. Results and Discussion: ) and projected area (R² = 0.96~0.99) have strong correlations with the manual measurement results. Additionally, 3D reconstruction results with different moving speeds and times throughout the day and in different scenes were also verified. The results show that the method can be applied to dynamic detection with a moving speed up to 0.6 m/s and can achieve acceptable detection results in the daytime, as well as at night. Thus, the proposed method can improve the efficiency of individual crop 3D point cloud data extraction with acceptable accuracy, which is a feasible solution for crop seedling 3D phenotyping outdoors.
Why it matches plant phenotyping methodsRGB-Dカメラと移動プラットフォームによる動的3D形質取得法を開発し、個体の草丈・葉面積・投影面積を手動測定と検証しており、植物フェノタイピング手法が中心である。
abstractwe propose a dynamic 3D data acquisition method in the field suitable for various crops by using a consumer-grade RGB-D camera installed on a ground-based movable platform
The dormant pruning of jujube is a labor-intensive and time-consuming activity in the production and management of jujube orchards, which mainly depends on manual operation. Automatic pruning using robots could be a better way to solve the shortage of skilled labor and improve efficiency. In order to realize automatic pruning of jujube trees, a method of pruning point identification based on skeleton information is presented. This study used an RGB-D camera to collect multi-view information on jujube trees and built a complete point cloud information model of jujube trees. The space colonization algorithm acts on the global point cloud to generate the skeleton of jujube trees. The iterative relationship between skeleton points was represented by constructing a directed graph. The proposed skeleton analysis algorithm marked the skeleton as the trunk, the primary branches, and the lateral branches and identified the pruning points under the guidance of pruning rules. Finally, the visual model of the pruned jujube tree was established through the skeleton information. The results showed that the registration errors of individual jujube trees were less than 0.91 cm, and the average registration error was 0.66 cm, which provided a favorable database for skeleton extraction. The skeleton structure extracted by the space colonization algorithm had a high degree of coincidence with jujube trees, and the identified pruning points were all located on the primary branches of jujube trees. The study provides a method to identify the pruning points of jujube trees and successfully verifies the validity of the pruning points, which can provide a reference for the location of the pruning points and visual research basis for automatic pruning.
Why it matches plant phenotyping methodsRGB-D点群からナツメ樹の樹幹・一次枝・側枝の骨格を抽出し、剪定点を推定する画像・計算手法が研究の中心であり、植物形態・樹体構造の表現型取得として妥当。
abstractThis study used an RGB-D camera to collect multi-view information on jujube trees and built a complete point cloud information model of jujube trees.
Highlights A commercial depth camera with a custom-designed graphical user interface was evaluated to detect tree canopy. Measurement variations under different indoor conditions were negligible for practical applications. Measurement errors ranged from 2.8% to 15.8%, which were acceptable for outdoor applications. Variation of crabapple canopy detection rate was less than 6% from sunrise to sunset. Abstract. To reduce crop protection product use and environmental impacts while maintaining application efficacy and convenience for applicators, an automatic variable rate sprayer coupled with a canopy detection sensor is required. A commercial depth camera was tested as a means of detecting the canopy of ornamental and tree crops for the sprayer. A custom-designed graphical user interface was developed to control the depth camera and save RGB and IR images and depth data to a local computer. Indoor evaluations showed that measurements could be influenced by the temperature and illumination; however, the influence was minimal, with a relative error of less than 1% and a maximum difference of 14 mm between the average measurements. The depth camera was able to detect a 31% to 72% area of a 20-mm wide target, and the rates went up 72% to 89% when the target width increased to 40 mm. The depth camera showed acceptable performance in detecting canopy contour changes and had measurement errors of 2.8% to 15.3% while detecting the distances to outdoor crabapple and oak trees. In addition, the depth camera detected tree canopy in various outdoor conditions from sunrise to sunset with reasonable accuracy (less than 10% of relative errors). In terms of measurement stability, the depth camera detected crabapple canopy with less than 6% variations under various illuminations between sunrise and sunset. The results suggested that the performance of the depth camera was adequate for detecting canopy under outdoor conditions for future variable-rate spray applications in ornamental and tree crop production. In addition, the study outlined the performance of the depth camera, which provided a guideline for future applications. Keywords: Machine Vision, Precision Agriculture, Specialty Crop, Stereo Vision, Variable Rate Application.
Why it matches plant phenotyping methods深度カメラによる樹冠検出・距離計測を中心に、屋内外で誤差や安定性を評価しており、植物形態の取得手法の技術検証に該当する。
abstractA commercial depth camera was tested as a means of detecting the canopy of ornamental and tree crops for the sprayer.
Orchard spraying robots must visually obtain citrus tree crown growth information to meet the variable growth-stage-based spraying requirements. However, the complex environments and growth characteristics of fruit trees affect the accuracy of crown segmentation. Therefore, we propose a feature-map-based squeeze-and-excitation UNet++ (MSEU) region-based convolutional neural network (R-CNN) citrus tree crown segmentation method that intakes red–green–blue-depth (RGB-D) images that are pixel aligned and visual distance-adjusted to eliminate noise. Our MSEU R-CNN achieves accurate crown segmentation using squeeze-and-excitation (SE) and UNet++. To fully fuse the feature map information, the SE block correlates image features and recalibrates their channel weights, and the UNet++ semantic segmentation branch replaces the original mask structure to maximize the interconnectivity between feature layers, achieving a near-real time detection speed of 5 fps. Its bounding box (bbox) and segmentation (seg) AP50 scores are 96.6 and 96.2%, respectively, and the bbox average recall and F1-score are 73.0 and 69.4%, which are 3.4, 2.4, 4.9, and 3.5% higher than the original model, respectively. Compared with bbox instant segmentation (BoxInst) and conditional convolutional frameworks (CondInst), the MSEU R-CNN provides better seg accuracy and speed than the previous-best Mask R-CNN. These results provide the means to accurately employ autonomous spraying robots.
Why it matches plant phenotyping methodsRGB-D画像から柑橘樹冠を抽出する画像解析手法を開発・評価しており、樹冠という植物形態・生育状態の取得が中心であるため、ロボット用途でも植物フェノタイピング手法に該当する。
abstractwe propose a feature-map-based squeeze-and-excitation UNet++ (MSEU) region-based convolutional neural network (R-CNN) citrus tree crown segmentation method that intakes red–green–blue-depth (RGB-D) images
Estimation of fruit size on-tree is useful for yield estimation, harvest timing and market planning. Automation of measurement of fruit size on-tree is possible using RGB-depth (RGB-D) cameras, if partly occluded fruit can be removed from consideration. An RGB-D Time of Flight camera was used in an imaging system that can be driven through an orchard. Three approaches were compared, being: (i) refined bounding box dimensions of a YOLO object detector; (ii) bounding box dimensions of an instance segmentation model (Mask R-CNN) applied to canopy images, and (iii) instance segmentation applied to extracted bounding boxes from a YOLO detection model. YOLO versions 3, 4 and 7 and their tiny variants were compared to an in-house variant, MangoYOLO, for this application, with YOLO v4-tiny adopted. Criteria developed to exclude occluded fruit by filtering based on depth, mask size, ellipse to mask area ratio and difference between refined bounding box height and ellipse major axis. The lowest root mean square error (RMSE) of 4.7 mm and 5.1 mm on the lineal length dimensions of a population (n = 104) of Honey Gold and Keitt varieties of mango fruit, respectively, and the lowest fruit exclusion rate was achieved using method (ii), while the RMSE on estimated fruit weight was 113 g on a population weight range between 180 and 1130 g. An example use is provided, with the method applied to video of an orchard row to produce a weight frequency distribution related to packing tray size.
Why it matches plant phenotyping methodsRGB-D画像と物体検出・インスタンスセグメンテーションを用いて樹上マンゴー果実のサイズ・重量を推定し、複数手法を比較検証しているため、植物表現型取得法が研究の中心である。
abstractThree approaches were compared, being: (i) refined bounding box dimensions of a YOLO object detector; (ii) bounding box dimensions of an instance segmentation model (Mask R-CNN) applied to canopy images, and (iii) instance segmentation applied to extracted bounding boxes from a YOLO detection model.
Reproduction assets foundThe paper publicly releases the RGB-D image datasets (Dataset-B and Dataset-C) used for training/testing the Mask R-CNN and YOLO-based mango fruit sizing models via a DOI deposit. No author analysis code or trained model checkpoints are explicitly deposited; the GitHub links cited are third-party frameworks (Darknet, MDataset · publicAll images in Dataset B and Dataset C used in
this study are available at https://doi.org/10.25946/21655628 (accessed on 15 October 2022).Open asset ↗10.25946/21655628pdf-page:4 lines:1-58Plant phenotyping relevance match · UnverifiedCrossref · OpenAlex · checked 8 Sept 2026
Pod phenotypic traits are closely related to grain yield and quality. Pod phenotype detection in soybean populations in natural environments is important to soybean breeding, cultivation, and field management. For an accurate pod phenotype description, a dynamic detection method is proposed based on an improved YOLO-v5 network. First, two varieties were taken as research objects. A self-developed field soybean three-dimensional color image acquisition vehicle was used to obtain RGB and depth images of soybean pods in the field. Second, the red–green–blue (RGB) and depth images were registered using an edge feature point alignment metric to accurately distinguish complex environmental backgrounds and establish a red–green–blue-depth (RGB-D) dataset for model training. Third, an improved feature pyramid network and path aggregation network (FPN+PAN) structure and a channel attention atrous spatial pyramid pooling (CA-ASPP) module were introduced to improve the dim and small pod target detection. Finally, a soybean pod quantity compensation model was established by analyzing the influence of the number of individual plants in the soybean population on the detection precision to statistically correct the predicted pod quantity. In the experimental phase, we analyzed the impact of different datasets on the model and the performance of different models on the same dataset under the same test conditions. The test results showed that compared with network models trained on the RGB dataset, the recall and precision of models trained on the RGB-D dataset increased by approximately 32% and 25%, respectively. Compared with YOLO-v5s, the precision of the improved YOLO-v5 increased by approximately 6%, reaching 88.14% precision for pod quantity detection with 200 plants in the soybean population. After model compensation, the mean relative errors between the predicted and actual pod quantities were 2% to 3% for the two soybean varieties. Thus, the proposed method can provide rapid and massive detection for pod phenotyping in soybean populations and a theoretical basis and technical knowledge for soybean breeding, scientific cultivation, and field management.
Why it matches plant phenotyping methods大豆莢数という植物形質を対象に、RGB-D画像取得、画像登録、改良YOLO-v5、補正モデルを開発・検証しており、フェノタイピング手法が研究の中心である。
abstractFor an accurate pod phenotype description, a dynamic detection method is proposed based on an improved YOLO-v5 network.
Volume is an important phenotype and quality attribute of sweetpotato storage roots. Conventionally the volume of most agricultural products is measured by water displacement. This method, which requires submerging the products in a container of water and measuring the displacement of water in the container, is time-consuming and tedious. It would be beneficial for sweetpotato breeding programs and quality inspection if a rapid method is developed for measuring the volume of sweetpotatoes. This study is therefore to evaluate the feasibility of LiDAR (light detection and ranging) technology as a novel high-throughput approach to phenotyping and measurement of the volume of sweetpotatoes. LiDAR data will be acquired from sweetpotato storage roots using a consumer-grade sensor, Intel® RealSense™ L515, which is an RGB-D (red-green-blue-depth) camera. Ground-truth volume values will be obtained using the reference water displacement method. RGB images will be used to segment sweetpotatoes from background, and extract meaningful features (e.g., the major axis length and the center of mass), complement the point cloud data from depth images for volume estimation. The shape of the sweetpotatoes will be constructed by a series of three-dimensional coordinate points, the alpha shape method is to be used to envelop the boundary points of sweetpotatoes to obtain a non-convex body, and thereby the volume of the sweet potato will be calculated. The efficacy of the proposed method will be evaluated in terms of volume estimation accuracy.
Why it matches plant phenotyping methodsLiDAR・RGB-D画像と画像セグメンテーション、3D形状再構成によりサツマイモ貯蔵根の体積を推定する表現型計測法を開発・評価しており、方法が研究の中心である。
abstractThis study is therefore to evaluate the feasibility of LiDAR (light detection and ranging) technology as a novel high-throughput approach to phenotyping and measurement of the volume of sweetpotatoes.
Field / plotLiDAR / point cloudRGB-D / ToFStem / branchCounting2D/3D reconstructionSegmentationArchitecture / morphology / geometry
Pruning is a time-consuming and labor-intensive practice for managing of dormant jujube orchards, in which dormant pruning is still mainly dependent on manual operation. Automated pruning using a robotic platform could be a better solution to overcome the skilled labor shortage and increased labor costs. With the development of dwarf and dense planting of jujube trees, the tree architectures are suitable for robotic pruning. This study concentrated on the detection and identification of pruning branches, which was a critical step for automating pruning operations. The method in this study mainly consisted of three steps: first, present a vision system based on two synchronous consumer-level RGB-D cameras to obtain the high-quality 3D point cloud in filed; second, propose the reconstruction pipeline for the desired three-dimensional model (basically a complete 3D model) using only two perspectives; and third, automatically segment the trunks and branches based on deep learning method (SPGNet), and then apply the DBSCAN clustering algorithm for estimating branch counts of jujube trees. For the reconstruction of jujube tree, registration errors II (0°-180°) were greater than errors I (0°-55°), and registration errors I (0°-55°) were less than 1 mm under different light conditions. Experimental results demonstrated that trunks and branches were segmented successfully with class accuracies of 0.93 and 0.84, and another metric intersection-over-union (IoU) was 0.85 and 0.76, respectively; the coefficient of determination (R²) values analyzed between the ground-truth and cluster results were 0.83, 0.88, and 0.89 in sunny, cloudy, and night conditions, respectively. These results showed that the proposed method for detecting branches could be utilized to generate substantial information, such as the diameter length and diameter of the branch, which was critical for the automated dormant pruning of jujube trees in the future orchard field.
Why it matches plant phenotyping methods3D画像再構成と深層学習によってナツメ樹の幹・枝を抽出し、枝数や直径などの植物構造形質を推定する手法を開発・評価しており、剪定ロボット向けの植物フェノタイピングが中心である。
abstractpresent a vision system based on two synchronous consumer-level RGB-D cameras to obtain the high-quality 3D point cloud in filed
Nitrogen is an important nutrient element for crop growth. A timely understanding of plant nitrogen information helps to adopt appropriate agricultural production management to maintain high yield and quality in wheat production. Crop spectral monitoring technology can obtain the leaf nitrogen content (LNC) information of wheat quickly and nondestructively. However, optical radiation interacts with the atmosphere, the canopy, and the soil before being captured by the sensor, and the capability to intercept, reflect, and transmit the radiation is different for different canopy structures. In wheat LNC monitoring, differences in target canopy structures will lead to changes in canopy reflectance, which can affect the monitoring accuracy. In this study, RGB and depth images were used to obtain wheat canopy structure indices. By analyzing the correlations between wheat LNC, canopy structure indices, and spectral reflectance, the indices that had large influence on spectral reflectance were screened; the change dynamics of these canopy structure indices under different internal and external factors were compared, and factors with greater impact were selected as the basis for grouping. On this basis, this study used the spectral indices RVI (660, 815) and RVI (730, 815) as fixed-effect variables and wheat population grouping variables as random-effect variables to construct linear mixed models of wheat LNC. In addition, the spectral indices and canopy structure indices were used as input parameters and wheat population grouping variables as output parameters to construct a random forest classifier for wheat canopy types. When monitoring the target population, we first predicted the classification of the unknown canopy by the classifier. The classification results and spectral indices were then input into the linear mixed models to realize the prediction of wheat LNC. The R² of the prediction for wheat LNC with RVI (660, 815) and RVI (730, 815) increased from 0.57 and 0.71 to 0.76 and 0.80, respectively, and the RRMSE decreased from 20.86% and 17.33% to 15.58% and 14.20%, respectively. This study used digital image information to compensate for spectral information, broadened the amount of information used for the remote sensing monitoring of farmland, and effectively improved the accuracy and universality of wheat LNC monitoring.
Why it matches plant phenotyping methodsRGB・深度画像とスペクトル情報を統合し、作物キャノピー構造を補償してコムギ葉窒素含量を推定する手法を構築・評価しており、フェノタイピング手法が研究の中心である。
abstractIn this study, RGB and depth images were used to obtain wheat canopy structure indices.
Plant height can be used for assessing plant vigor and predicting biomass and yield. Manual measurement of plant height is time-consuming and labor-intensive. We describe a method for measuring maize plant height using an RGB-D camera that captures a color image and depth information of plants under field conditions. The color image was first processed to locate its central area using the S component in HSV color space and the Density-Based Spatial Clustering of Applications with Noise algorithm. Testing showed that the central areas of plants could be accurately located. The point cloud data were then clustered and the plant was extracted based on the located central area. The point cloud data were further processed to generate skeletons, whose end points were detected and used to extract the highest points of the central leaves. Finally, the height differences between the ground and the highest points of the central leaves were calculated to determine plant heights. The coefficients of determination for plant heights manually measured and estimated by the proposed approach were all greater than 0.95. The method can effectively extract the plant from overlapping leaves and estimate its plant height. The proposed method may facilitate maize height measurement and monitoring under field conditions.
Why it matches plant phenotyping methodsRGB-D画像と点群処理を用いて圃場のトウモロコシ草丈を推定する手法を開発・検証しており、表現型取得が研究の中心である。
abstractWe describe a method for measuring maize plant height using an RGB-D camera that captures a color image and depth information of plants under field conditions.
A large number of apple seedlings are planted in orchards each year, where accurate and fast seedling grading to ensure their quality before planting has become a crucial problem. However, seedling grading by manual measurement of morphological indicators is laborious and inaccurate, and it’s thus highly desirable to be replaced by machine vision. Seedling segmentation is one of the key steps of measuring morphological indicators and grading by machine vision. Therefore, a segmentation method of apple seedlings based on BlendMask with ResNet-101 to do transfer learning was proposed. A total of 450 original images were captured with Azure Kinect DK sensor. Root, rootstock, graft union, and scion of apple seedlings were labeled using a novel labeling strategy, which probably affect segmentation of thin and long objects. Scion was labeled with three different strategies, namely whole labeling (WL), segmental labeling (SL), and segmental-end-merge labeling (SEML). Results showed that the most suitable strategy was the SL for scion, which obtained a mean average precision of 91.2 % and the highest mean intersection over union of 79.3 % in the three labeling strategies. The average precisions of root, rootstock, graft union, and scion with the SL were 98.9 %, 89.3 %, 90.6 %, and 85.6 %, respectively. Intersection over unions of root, rootstock, graft union, and scion by the SL were 87.2 %, 75.8 %, 69.3 %, and 84.9 %, respectively. And it cost about 285 ms on average to process an image with resolution 3840 × 2160 pixels. The above results illustrated that the SL strategy is conducive to improve segmentation precision of thin and long objects. Moreover, apple seedlings can be effectively segmented, which is beneficial for the machine vision to measure morphological indicators and grade.
Why it matches plant phenotyping methodsリンゴ苗の形態指標測定と等級付けに用いる画像セグメンテーション手法を開発・評価しており、植物フェノタイピング手法が中心である。
abstractTherefore, a segmentation method of apple seedlings based on BlendMask with ResNet-101 to do transfer learning was proposed.
Active perception for fruit mapping and harvesting is a difficult task since occlusions occur frequently and the location as well as size of fruits change over time. State-of-the-art viewpoint planning approaches utilize computationally expensive ray casting operations to find good viewpoints aiming at maximizing information gain and covering the fruits in the scene. In this paper, we present a novel viewpoint planning approach that explicitly uses information about the predicted fruit shapes to compute targeted viewpoints that observe as yet unobserved parts of the fruits. Furthermore, we formulate the concept of viewpoint dissimilarity to reduce the sampling space for more efficient selection of useful, dissimilar viewpoints. Our simulation experiments with a UR5e arm equipped with an RGB-D sensor provide a quantitative demonstration of the efficacy of our iterative next best view planning method based on shape completion. In comparative experiments with a state-of-the-art viewpoint planner, we demonstrate improvement not only in the estimation of the fruit sizes, but also in their reconstruction, while significantly reducing the planning time. Finally, we show the viability of our approach for mapping sweet peppers plants with a real robotic system in a commercial glasshouse.
Why it matches plant phenotyping methods果実形状の再構成とサイズ推定を目的とする視点計画法を開発・比較評価しており、植物表現型取得が中心的な技術貢献である。
abstractwe present a novel viewpoint planning approach that explicitly uses information about the predicted fruit shapes to compute targeted viewpoints that observe as yet unobserved parts of the fruits.
Field / plotLiDAR / point cloudRGB-D / ToFFruitMorphology / geometry measurementObject detectionSegmentationFruit / seed / panicle traits
Fruit detection and localization is of great significance for horticulture work and robotic harvesting in orchards. Although the existing studies of fruit detection have achieved good results based on 2D image analysis, accurate fruit detection on trees is still challenging because of illumination changes, shielding of leaves and branches, overlapping of fruits and so on. To improve the accuracy of fruit detection and location, this paper proposes a novel ripe pomegranate fruit detection and location method based on improved F-PointNet and 3D clustering method, which is consisting of: (1) RGB-D feature fusion Mask R-CNN was used to realize fruit detection and segmentation; (2) PointNet combined with OPTICS algorithm based on manifold distance and PointFusion was used to segment point clouds in the frustum fruit region, and 3D box was placed in the region of interest; (3) The sphere fitting was performed to obtain the position and the size of a pomegranate. The comparative experiments have been carried out and analyzed, the RGB-D feature fusion Mask R-CNN has the best performance with the F1 score of 0.845 and the AP score of 0.952 respectively, and the improved F-PointNet has better performance than the classical F-PointNet. The measurement radius experiment results of 100 pomegranate samples randomly selected demonstrate that the RMSE is 0.235 cm, the R² is 0.826, and the position error is less than 5 mm. These results validate that the proposed detection and location method can effectively detect and locate a single ripe pomegranate under unstructured orchard environment.
Why it matches plant phenotyping methods果実の検出・位置推定に加え、球フィッティングでザクロの位置とサイズを定量化する画像・3D計測手法を開発し、比較実験と精度検証を行っており、植物表現型取得が中心である。
abstractthis paper proposes a novel ripe pomegranate fruit detection and location method based on improved F-PointNet and 3D clustering method
In this study, we proposed a system to quantify the plant height from a three-dimensional point cloud of a plant using the INTEL REALSENSE DEPTH CAMERA D415. This system was used to compare and evaluate plant growth under LEP, LED, and fluorescent lighting. In the experiment, four basil plants were hydroponically cultivated as measurement targets. The highest point from the three-dimensional point cloud of each plant reconstructed by D415 was calculated as the plant height, and the change for about one month was recorded. As a result of the experiment, it was confirmed that the plants grow faster in the order of LEP, LED, and fluorescent lamp. Regarding the illuminance, LEP, fluorescent lamp, and LED were the highest in this order. LEP is advantageous for plant growth because it can emit light with a wide band spectrum with high efficiency. Cultivation using LED tended to grow faster than fluorescent lamps. Since the LED emits only light in the wavelength band related to plant growth, it was considered that efficient plant growth was realized. It was confirmed that the proposed measurement system not only enables easy quantification of plant height, but also enables visualization of growth by presenting a three-dimensional point cloud. Under artificial light illumination, it became clear that the accuracy of 3D reconstruction differs between when artificial light is on and when it is off. Under artificial light lighting, the reconstruction accuracy tends to be lower, and improvement is a future task.
Why it matches plant phenotyping methods3D点群から植物高を定量化する測定システムの提案と、照明条件下での再構成精度評価が研究の中心であり、植物表現型取得手法に該当する。
abstractwe proposed a system to quantify the plant height from a three-dimensional point cloud of a plant using the INTEL REALSENSE DEPTH CAMERA D415.
Growth traits, such as fresh weight, diameter, and leaf area, are pivotal indicators of growth status and the basis for the quality evaluation of lettuce. The time-consuming, laborious and inefficient method of manually measuring the traits of lettuce is still the mainstream. In this study, a three-stage multi-branch self-correcting trait estimation network (TMSCNet) for RGB and depth images of lettuce was proposed. The TMSCNet consisted of five models, of which two master models were used to preliminarily estimate the fresh weight (FW), dry weight (DW), height (H), diameter (D), and leaf area (LA) of lettuce, and three auxiliary models realized the automatic correction of the preliminary estimation results. To compare the performance, typical convolutional neural networks (CNNs) widely adopted in botany research were used. The results showed that the estimated values of the TMSCNet fitted the measurements well, with coefficient of determination ( R 2 ) values of 0.9514, 0.9696, 0.9129, 0.8481, and 0.9495, normalized root mean square error (NRMSE) values of 15.63, 11.80, 11.40, 10.18, and 14.65% and normalized mean squared error (NMSE) value of 0.0826, which was superior to compared methods. Compared with previous studies on the estimation of lettuce traits, the performance of the TMSCNet was still better. The proposed method not only fully considered the correlation between different traits and designed a novel self-correcting structure based on this but also studied more lettuce traits than previous studies. The results indicated that the TMSCNet is an effective method to estimate the lettuce traits and will be extended to the high-throughput situation. Code is available at https://github.com/lxsfight/TMSCNet.git.
Why it matches plant phenotyping methodsRGB・深度画像からレタスの複数形質を推定する新規ネットワークを開発し、既存手法と性能比較しており、植物フェノタイピング手法が研究の中心である。
abstracta three-stage multi-branch self-correcting trait estimation network (TMSCNet) for RGB and depth images of lettuce was proposed
Reproduction assets foundThe paper uses the public Autonomous Greenhouses Challenge 3 dataset (RGB/depth lettuce images with FW/DW/H/D/LA measurements) and states author code availability on GitHub.Code · publicCode is available at https://github.com/lxsfight/TMSCNet.git .Open asset ↗github.com/lxsfight/TMSCNetlines:1-41Code / dataset availability confirmedEurope PMC · OpenAlex · checked 8 Sept 2026
Fresh weight is a widely used growth indicator for quantifying crop growth. Traditional fresh weight measurement methods are time-consuming, laborious, and destructive. Non-destructive measurement of crop fresh weight is urgently needed in plant factories with high environment controllability. In this study, we proposed a multi-modal fusion based deep learning model for automatic estimation of lettuce shoot fresh weight by utilizing RGB-D images. The model combined geometric traits from empirical feature extraction and deep neural features from CNN. A lettuce leaf segmentation network based on U-Net was trained for extracting leaf boundary and geometric traits. A multi-branch regression network was performed to estimate fresh weight by fusing color, depth, and geometric features. The leaf segmentation model reported a reliable performance with a mIoU of 0.982 and an accuracy of 0.998. A total of 10 geometric traits were defined to describe the structure of the lettuce canopy from segmented images. The fresh weight estimation results showed that the proposed multi-modal fusion model significantly improved the accuracy of lettuce shoot fresh weight in different growth periods compared with baseline models. The model yielded a root mean square error (RMSE) of 25.3 g and a coefficient of determination ( R 2 ) of 0.938 over the entire lettuce growth period. The experiment results demonstrated that the multi-modal fusion method could improve the fresh weight estimation performance by leveraging the advantages of empirical geometric traits and deep neural features simultaneously.
Why it matches plant phenotyping methodsRGB-D画像からレタスの生体重を非破壊推定する画像解析・深層学習手法の開発が研究の中心であり、植物表現型取得法に該当する。
abstractA lettuce leaf segmentation network based on U-Net was trained for extracting leaf boundary and geometric traits.
Reproduction assets foundThe paper's phenotyping inputs (top-view RGB and aligned depth images of 388 lettuces with destructively measured traits) come from the publicly available 3rd Autonomous Greenhouse Challenge Online Challenge Lettuce Images dataset, with an explicit public URL in the data availability statement. No author analysis code,Dataset · publicPublicly available datasets were analyzed in this study. This data can be found here: https://data.4tu.nl/articles/dataset/3rd_Autonomous_Greenhouse_Challenge_Online_Challenge_Lettuce_Images/15023088 .Open asset ↗data.4tu.nl · 15023088lines:657-691Plant phenotyping relevance match · UnverifiedEurope PMC · OpenAlex · checked 8 Sept 2026
Abstract Plant shape and structure are important factors in peanut breeding research. Constructing a three-dimension (3D) model can provide an effective digital tool for comprehensive and quantitative analysis of peanut plant structure. A 3D reconstruction method based on dual RGB-D cameras was proposed for the peanut plant 3D model quickly and accurately. The two Kinect v2 were mirror symmetry placed on both sides of the peanut plant, and the point cloud data obtained were filtered twice to remove noise interference. After rotation and translation based on the corresponding geometric relationship, the point cloud acquired by the two Kinect v2 was converted to the same coordinate system and spliced into the 3D structure of the peanut plant. The experiment was conducted at various growth stages based on twenty potted peanuts. The plant traits’ height, width, length, and volume were calculated through the reconstructed 3D models, and manual measurement was carried out at the same time. The accuracy of the 3D model was evaluated through a synthetic coefficient, which was generated by calculating the average accuracy of the four traits. The test result shows that the synthetic accuracy of the reconstructed peanut plant 3D model by this method is 93.42%. A comparative experiment with the iterative closest point (ICP) algorithm, a widely used 3D modeling algorithm, was additionally implemented to test the rapidity of this method. The test result shows that the proposed method is 2.54 times faster with approximated accuracy compared to the ICP method. This approach should be useful for 3D modeling and phenotyping peanut breeding.
Why it matches plant phenotyping methodsピーナッツの3D形状・構造をRGB-Dカメラで再構成し、複数の植物形質を推定・精度検証する手法開発が研究の中心であるため。
abstractA 3D reconstruction method based on dual RGB-D cameras was proposed for the peanut plant 3D model quickly and accurately.
Apple tree phenotyping can reflect individual development of single apple tree, which mainly involves tree height, crown width, and diameter of apple tree trunk (DATT). This study aimed to estimate diameter of grafted apple tree trunk, whose target position of diameter estimation is about 10 cm above grafting position. An estimated DATT approach of combining red–greenblue-depth (RGB-D) sensor with SOLOv2 was proposed. Firstly, Kinect V2 was employed to obtain original RGB images and point clouds of the grafted apple trees simultaneously. There were 120 and 60 RGB images and corresponding point clouds randomly collected from two modern apple orchards. Secondly, SOLOv2 deep learning model was selected and trained to instance segment grafting position from RGB image for determining it automatically. Then, corresponding exact position of the grafting position in point cloud was mapped by coordinate transformation of its pixel coordinates, which was obtained by trained SOLOv2 model. Finally, DATT was estimated by calculating the difference between maximum and minimum Y coordinates of points selected by distance thresholds in X, Y, and Z directions near the target position, which were 0.10 m, 0.035 m, and 0.20 m, respectively. Results showed that average precision and average recall of the trained SOLOv2 model for instant segmenting the grafting position were 0.811 and 0.830, respectively. Mean absolute error, mean absolute percentage error, and root mean square error of the proposed method were 3.01 mm, 5.86%, and 3.79 mm, respectively. It illustrates that the proposed method can estimate DATT and thus contribute to automatic apple tree phenotyping.
Why it matches plant phenotyping methodsRGB-Dセンサ、点群、インスタンスセグメンテーションを組み合わせ、リンゴ樹幹径という植物形質を推定する手法を開発・検証しており、フェノタイピング手法が中心である。
abstractAn estimated DATT approach of combining red–greenblue-depth (RGB-D) sensor with SOLOv2 was proposed.
Growth indices can quantify crop productivity and establish optimal environmental, nutritional, and irrigation control strategies. A convolutional neural network (CNN)-based model is presented for estimating various growth indices (i.e., fresh weight, dry weight, height, leaf area, and diameter) of four varieties of greenhouse lettuce using red, green, blue, and depth (RGB-D) data obtained using a stereo camera. Data from an online autonomous greenhouse challenge (Wageningen University, June 2021) were employed in this study. The data were collected using an Intel RealSense D415 camera. The developed model has a two-stage CNN architecture based on ResNet50V2 layers. The developed model provided coefficients of determination from 0.88 to 0.95, with normalized root mean square errors of 6.09%, 6.30%, 7.65%, 7.92%, and 5.62% for fresh weight, dry weight, height, diameter, and leaf area, respectively, on unknown lettuce images. Using red, green, blue (RGB) and depth data employed in the CNN improved the determination accuracy for all five lettuce growth indices due to the ability of the stereo camera to extract height information on lettuce. The average time for processing each lettuce image using the developed CNN model run on a Jetson SUB mini-PC with a Jetson Xavier NX was 0.83 s, indicating the potential for the model in fast real-time sensing of lettuce growth indices.
Why it matches plant phenotyping methodsRGB-D画像とCNNを用いてレタスの複数の生育形質を推定する手法を開発・検証しており、表現型取得が研究の中心である。
abstractA convolutional neural network (CNN)-based model is presented for estimating various growth indices (i.e., fresh weight, dry weight, height, leaf area, and diameter) of four varieties of greenhouse lettuce using red, green, blue, and depth (RGB-D) data obtained using a stereo camera.
Reproduction assets foundThe paper's phenotyping inputs (388 RGB-D lettuce image pairs with destructive growth-index measurements from the Third Autonomous Greenhouse Challenge) are a third-party public dataset explicitly stated to be publicly available at 4TU.ResearchData, with the DOI 10.4121/15023088.v1 cited in the text and figure captionsDataset · publicThe dataset is available in online: https://doi.org/10.4121/15023088.v1 [ 30 ].Open asset ↗10.4121/15023088.v1lines:518-697Plant phenotyping relevance match · UnverifiedOpenAlex · checked 15 Sept 2026
Monitoring plants and fruits is important in modern agriculture, with applications ranging from high-throughput phenotyping to autonomous harvesting. Obtaining highly accurate 3D measurements under real agricultural conditions is a challenging task. In this letter, we address the problem of estimating the 3D shape of fruits when only a partial view is available. We propose a pipeline that exploits high-resolution 3D data in the learning phase but only requires a single RGB-D frame to predict the 3D shape of acompletefruit during operation. To achieve this, we first learn a latent space of potential fruit appearances that we can decode into an SDF volume. With the pretrained, frozen decoder, we subsequently learn an encoder that can produce meaningful latent vectors from a single RGB-D frame. The experiments presented in this letter suggest that our approach can predict the 3D shape of whole fruits online, needing only 4 ms for inference. We evaluate our approach in controlled environments and illustrate its deployment in greenhouses without modifications.
Why it matches plant phenotyping methodsRGB-D画像から果実全体の3D形状を推定する手法を開発しており、収穫対象の単なる検出ではなく、再利用可能な果実形状という植物器官形質を定量化している。
abstractwe address the problem of estimating the 3D shape of fruits when only a partial view is available.
Tomato sucker or axillary shoots should be removed to increase the yield and reduce the disease on tomato plants. It is an essential step in the tomato plant care process. It is usually performed manually by farmers. An automated approach can save a lot of time and labor. In the literature review, we see that semantic segmentation is a process of recognizing or classifying each pixel in an image, and it can help machines recognize and localize tomato suckers. This paper proposes a semantic segmentation neural network that can detect tomato suckers quickly by the tomato plant images. We choose RGB-D images which capture not only the visual of objects but also the distance information from objects to the camera. We make a tomato RGB-D image dataset for training and evaluating the proposed neural network. The proposed semantic segmentation neural network can run in real-time at 138.2 frames per second. Its number of parameters is 680, 760, much smaller than other semantic segmentation neural networks. It can correctly detect suckers at 80.2%. It requires low system resources and is suitable for the tomato dataset. We compare it to other popular non-real-time and real-time networks on the accuracy, time of execution, and sucker detection to prove its better performance.
Why it matches plant phenotyping methodsトマトの腋芽をRGB-D画像からセマンティックセグメンテーションで検出・局在化する手法を開発し、データセット作成と性能比較・検証を行っており、植物形態の取得が中心である。
abstractThis paper proposes a semantic segmentation neural network that can detect tomato suckers quickly by the tomato plant images.
In the process of farm automation, fruit detection is the basis and guarantee for yield prediction, automatic picking, and other orchard operations. RGB images can only obtain the two-dimensional information of the scene, which is not sufficient to effectively distinguish fruits that are dense growth and occlusion by branches and leaves. With the development of depth sensors, using RGB-D images with more complementary information can boost the performance of fruit detection. However, due to the nature of sensors and scene configurations, the quality of outdoor depth images is poor, posing a challenge when fusing RGB-D features. Therefore, this paper proposes an end-to-end RGB-D object detection network, termed as noise-tolerant feature fusion network (NT-FFN), to utilize the outdoor multi-modal data properly and improve the detection accuracy. Specifically, the NT-FFN first uses two structurally identical feature extractors to extract single-modal (color and depth) features, which is the base of the subsequent feature fusion. Then, to avoid introducing too much depth noise and focus the perception on the important part of the features, an attention-based fusion module is designed to adaptively fuse the multi-modal features. Finally, multi-scale features from the color images and the fusion modules are used to predict object position, which not only improves the network's ability to detect multi-scale objects but also further enhances the noise immunity of the network. In addition, this paper constructs an RGB-D citrus fruit dataset, which contributes to comprehensively evaluating the proposed network. Evaluation metrics on the dataset show that the NT-FFN achieves an AP⁵⁰ of 95.4% with a real-time speed, which outperforms single-modal methods, common multi-modal fusion strategies, and advanced multi-modal detection methods. The proposed NT-FFN also achieves excellent detection results in other fruit detection tasks, which verifies its generalization ability. This study provides the possibility and foundation for performing multi-modal information fusion in outdoor fruit detection.
Why it matches plant phenotyping methodsRGB-D画像から果実の位置を検出する手法を開発し、独自データセットで評価しているため、植物器官の観測・抽出方法が中心的である。
abstractTherefore, this paper proposes an end-to-end RGB-D object detection network, termed as noise-tolerant feature fusion network (NT-FFN), to utilize the outdoor multi-modal data properly and improve the detection accuracy.
Seedling transplantation is the key link in the automation of protected horticulture, and the transplantation quality will directly affect the crop yield. In the previous research, our research group developed a prototype of a low-loss transplantation robot for plug seedlings based on machine vision. On this basis, a method of planning seedling path by integrating seedling edge recognition technology and end effector of transplanting manipulator is proposed to reduce the loss of seedling transplantation. Firstly, the RGB image and depth image of the whole row of seedlings were obtained from the side of the plug seedling with Intel RealSense D415 depth camera. Then, the extreme point coordinates E (Xcb, Ycₐ) of seedling edge were obtained by edge recognition algorithm, and the coordinate value of z-axis was known. Taking point E as the starting point the transplanting manipulator picked up seedlings from the E point in an “L” path, which can reduce the damage of transplanting manipulator to seedling stems and leaves during seedling taking, thereby reducing the transplanting loss. Through three-factor four-level orthogonal test, it was determined that the distance between the depth camera and the plug tray was 600 mm, the height from the horizontal plane was 135 mm, and the light intensity was level 7. Then, a calibration test of extreme points on the edge of plug seedlings was carried out, and the calibration success rate was 98.4%. The deviation of X coordinate was within 5 mm, and the average ratio of deviation was 12.8%. The deviation of Y coordinate was within 4 mm, and the average ratio of deviation was 3%. The quality of image and the accuracy of depth information were the main causes of the deviation. Then, the comparative experiments of routine transplantation group, fixed transplantation group and machine vision transplantation group were carried out. The results showed that compared with the conventional transplantation group, the injury rate of machine vision group decreased by 11.11%, and the average time of single transplantation increased by 0.029 s. Compared with the fixed transplantation group, the injury rate increased by 0.46%, and the average transplantation time decreased by 0.238 s. Because the edge recognition time and the transplantation time intersected and did not interfere, only the time for the first seedling edge recognition was increased in the pipeline operation. Therefore, this method can reduce the damage rate while ensuring the efficiency. This study can provide a reference for reducing the injury rate of leafy vegetable seedlings transplantation.
Why it matches plant phenotyping methods深度カメラによる苗のエッジ認識・座標抽出を中核として、移植ロボットの経路計画と認識精度を検証しており、植物形態の取得手法が実質的な技術貢献である。
abstracta method of planning seedling path by integrating seedling edge recognition technology and end effector of transplanting manipulator is proposed
In agriculture, the majority of vision systems perform still image classification. Yet, recent work has highlighted the potential of spatial and temporal cues as a rich source of information to improve the classification performance. In this paper, we propose novel approaches to explicitly capture both spatial and temporal information to improve the classification of deep convolutional neural networks. We leverage available RGB-D images and robot odometry to perform inter-frame feature map spatial registration. This information is then fused within recurrent deep learnt models, to improve their accuracy and robustness. We demonstrate that this can considerably improve the classification performance with our best performing spatial-temporal model (ST-Atte) achieving absolute performance improvements for intersection-over-union (IoU[%]) of 4.7 for crop-weed segmentation and 2.6 for fruit (sweet pepper) segmentation. Furthermore, we show that these approaches are robust to variable framerates and odometry errors, which are frequently observed in real-world applications.
Why it matches plant phenotyping methodsRGB-D画像とロボットオドメトリを用いて空間・時間情報を統合する画像解析手法を開発し、作物・果実のセグメンテーション性能を検証しているため、植物表現型取得手法が中心である。
abstractWe demonstrate that this can considerably improve the classification performance with our best performing spatial-temporal model (ST-Atte) achieving absolute performance improvements for intersection-over-union (IoU[%]) of 4.7 for crop-weed segmentation and 2.6 for fruit (sweet pepper) segmentation.
TomatoGreenhouseRGB-D / ToFStem / branchSegmentationGrowth / development / phenology
Background On tomato plants, the flowering truss is a group or cluster of smaller stems where flowers and fruit develop, while the growing truss is the most extended part of the stem. Because the state of the growing truss reacts sensitively to the surrounding environment, it is essential to control its growth in the early stages. With the recent development of information and artificial intelligence technology in agriculture, a previous study developed a real-time acquisition and evaluation method for images using robots. Furthermore, we used image processing to locate the growing truss to extract growth information. Among the different vision algorithms, the CycleGAN algorithm was used to generate and transform unpaired images using generated learning images. In this study, we developed a robot-based system for simultaneously acquiring RGB and depth images of the growing truss of the tomato plant. Results The segmentation performance for approximately 35 samples was compared via false negative (FN) and false positive (FP) indicators. For the depth camera image, we obtained FN and FP values of 17.55 ± 3.01% and 17.76 ± 3.55%, respectively. For the CycleGAN algorithm, we obtained FN and FP values of 19.24 ± 1.45% and 18.24 ± 1.54%, respectively. When segmentation was performed via image processing through depth image and CycleGAN, the mean intersection over union (mIoU) was 63.56 ± 8.44% and 69.25 ± 4.42%, respectively, indicating that the CycleGAN algorithm can identify the desired growing truss of the tomato plant with high precision. Conclusions The on-site possibility of the image extraction technique using CycleGAN was confirmed when the image scanning robot drove in a straight line through a tomato greenhouse. In the future, the proposed approach is expected to be used in vision technology to scan tomato growth indicators in greenhouses using an unmanned robot platform.
Why it matches plant phenotyping methodsトマトの生育器官である生長トラスを対象に、RGB・深度画像、ロボット、CycleGANによる抽出・セグメンテーション手法を開発し、FN・FP・mIoUで性能評価しているため、植物表現型取得が中心である。
abstractwe developed a robot-based system for simultaneously acquiring RGB and depth images of the growing truss of the tomato plant.
In the process of green apple harvesting or yield estimation, affected by the factors, such as fruit color, light, and orchard environment, the accurate recognition and fast location of the target fruit brings tremendous challenges to the vision system. In this article, we improve a density peak cluster segmentation algorithm for RGB images with the help of a gradient field of depth images to locate and recognize target fruit. Specifically, the image depth information is adopted to analyze the gradient field of the target image. The vorticity center and two-dimensional plane projection are constructed to realize the accurate center location. Next, an optimized density peak clustering algorithm is applied to segment the target image, where a kernel density estimation is utilized to optimize the segmentation algorithm, and a double sort algorithm is applied to efficiently obtain the accurate segmentation area of the target image. Finally, the segmentation area with the circle center is the target fruit area, and the maximum value method is employed to determine the radius. The above two results are merged to achieve the contour fitting of the target fruits. The novel method is designed without iteration, classifier, and several samples, which has greatly improved operating efficiency. The experimental results show that the presented method significantly improves accuracy and efficiency. Meanwhile, this new method deserves further promotion.
Why it matches plant phenotyping methodsRGB-D画像からリンゴ果実を認識・位置推定し、セグメンテーションと輪郭抽出を行う手法が研究の中心であり、果実の形態・位置という植物形質を直接推定しているため。
abstractwe improve a density peak cluster segmentation algorithm for RGB images with the help of a gradient field of depth images to locate and recognize target fruit.
In the field of computer vision, 3D reconstruction of crops plays a crucially important role in agriculture. On-ground assessment of geometrical features of vineyards is of vital importance to generate valuable information that enables producers to take the optimum actions in terms of agricultural management. A training system of vines (Vitis vinifera L.), which involves pruning and a trellis system, results in a particular vine architecture, which is vital throughout the phenological stages. Pruning is required to maintain the vine’s health and to keep its productivity under control. The creation of 3D models of vineshoots is of crucial importance for management planning. Volume and structural information can improve pruning systems, which can increase crop yield and improve crop management. In this experiment, an RGB-D camera system, namely Kinect v2, was used to reconstruct 3D vine models, which were used to determine shoot volume on eight differentiated vineyard training systems: Lyre, GDC (Geneva Double Curtain), Y-Trellis, Pergola, Single Curtain, Smart Dyson, VSP (Vertical Shoot Positioned), and the head-trained Gobelet. The results were compared with dry biomass ground truth-values. Dense point clouds had a substantial impact on the connection between the actual biomass measurements in four of the training systems (Pergola, Curtain, Smart Dyson and VSP). For the comparison of actual dry biomass and RGB-D volume and its associated 3D points, strong linear fits were obtained. Significant coefficients of determination (R2 = 0.72 to R2 = 0.88) were observed according to the number of points connected to each training system separately, and the results revealed good correlations with actual biomass and volume values. When comparing RGB-D volume to weight, Pearson’s correlation coefficient increased to 0.92. The results reveal that the RGB-D approach is also suitable for shoot reconstruction. The research proved how an inexpensive optical sensor can be employed for rapid and reproducible 3D reconstruction of vine vegetation that can improve cultural practices such as pruning, canopy management and harvest.
Why it matches plant phenotyping methodsRGB-D画像からブドウ樹の3D構造・シュート体積を推定し、乾物バイオマスと検証する手法が研究の中心である。
abstractan RGB-D camera system, namely Kinect v2, was used to reconstruct 3D vine models, which were used to determine shoot volume
Modern consumer-grade RGB-D cameras provide intensive depth estimation and high frame rates. They have been widely used in agriculture. However, depth anamorphose occurs when using the RGB-D camera. In order to address this issue, this paper proposes a novel approach to correct the distorted depth images that fit the relationship between the true distances and the depth values from depth images. This study considers the structured light camera ASUS Xtion PRO LIVE as example to develop a system for obtaining a series of depth images at different distances from the target plane. A comparison analysis is conducted between the images before and after correction to evaluate the performance of the distortion correction. The point cloud image of corrected maize leaves is well coincident with the original plant. This approach improves the accuracy of depth measurement and optimizes the subsequent use of the depth camera in crop reconstruction and phenotyping studies.
Why it matches plant phenotyping methods深度カメラの歪み補正法を開発・比較評価し、トウモロコシ葉の3D再構成と表現型計測への利用精度を改善する研究であり、植物表現型取得手法が中心である。
abstractthis paper proposes a novel approach to correct the distorted depth images
Published30 May 2022The international archives of the photogrammetry, remote sensing and spatial information sciences/International archives of the photogrammetry, remote sensing and spatial information sciencesCited by 5 · OpenAlex ↗
Abstract. In the last few years, a number of low-cost 3D scanning sensors have been developed to reconstruct the real-world environment. These sensors were primarily designed for indoor use, making them highly unpredictable in terms of their performance and accuracy when used outdoors. The Azure Kinect belongs to this category of low-cost 3D scanners and has been successfully employed in outdoor applications. In addition, this sensor possesses features such as portability and live visualization during data acquisition that makes it extremely interesting in the field of forestry. In the context of forest inventory, these advantages would allow to facilitate the task of tree parameters acquisition in an efficient manner. In this paper, a protocol was established for the acquisition of 3D data in forests using the Azure Kinect. A comparison of the resulting point cloud was performed against photogrammetry. Results demonstrated that the Azure Kinect point cloud was of suitable quality for extracting tree parameters such as diameter at breast height (DBH, with a standard deviation of 2.2cm). Furthermore, the quality of the visual and geometric information of the point cloud was evaluated in terms of its feasibility to identify microhabitats. Microhabitats represent valuable information on forest biodiversity and are included in Swiss forest inventory measurements. In total, five different microhabitats were identified in the Azure Kinect Point cloud. The measurements were therefore comparable to sensors such as terrestrial laser scanning and photogrammetry. Therefore, we argue that the Azure Kinect point cloud can efficiently identify certain types of microhabitats and this study presents a first approach of its application in forest inventories.
Why it matches plant phenotyping methodsAzure Kinectによる森林内の3Dデータ取得プロトコルを確立し、写真測量と比較検証して、胸高直径や樹木マイクロハビタットを抽出する方法が中心である。
abstractIn this paper, a protocol was established for the acquisition of 3D data in forests using the Azure Kinect.
Plant breeders, scientists, and commercial producers commonly use growth rate as an integrated signal of crop productivity and stress. Plant growth monitoring is often done destructively via growth rate estimation by harvesting plants at different growth stages and simply weighing each individual plant. Within plant breeding and research applications, and more recently in commercial applications, non-destructive growth monitoring is done using computer vision to segment plants in images from the background, either in 2D or 3D, and relating these image-based features to destructive biomass measurements. Recent advancements in machine learning have improved image-based localization and detection of plants, but such techniques are not well suited to make biomass predictions when there is significant self-occlusion or occlusion from neighboring plants, such as those encountered under leafy green production in controlled environment agriculture. To enable prediction of plant biomass under occluded growing conditions, we develop an end-to-end deep learning approach that directly predicts lettuce plant biomass from color and depth image data as provided by a low cost and commercially available sensor. We test the performance of the proposed deep neural network for lettuce production, observing a mean prediction error of 7.3% on a comprehensive test dataset of 864 individuals and substantially outperforming previous work on plant biomass estimation. The modeling approach is robust to the busy and occluded scenes often found in commercial leafy green production and requires only measured mass values for training. We then demonstrate that this level of prediction accuracy allows for rapid, non-destructive detection of changes in biomass accumulation due to experimentally induced stress induction in as little as 2 days. Using this method growers may observe and react to changes in plant-environment interactions in near real time. Moreover, we expect that such a sensitive technique for non-destructive biomass estimation will enable novel research and breeding of improved productivity and yield in response to stress.
Why it matches plant phenotyping methodsRGB-D画像と深層学習を用いて、遮蔽下のレタス個体バイオマスを非破壊推定する手法を開発・評価しており、植物表現型取得が研究の中心です。
abstractwe develop an end-to-end deep learning approach that directly predicts lettuce plant biomass from color and depth image data as provided by a low cost and commercially available sensor.
Reproduction assets foundThe article provides an authors' public GitHub repository containing the analysis code for the RGB-D deep learning biomass estimation pipeline. The raw image/biomass dataset is only available on request (no public deposit).Code · publicCode available at https://github.com/NicoBux/Plant-Biomass-Monitoring .Open asset ↗NicoBux/Plant-Biomass-Monitoringlines:466-524Plant phenotyping relevance match · UnverifiedEurope PMC · checked 15 Sept 2026
Published1 Apr 2022Computers and Electronics in Agriculture.
Non-destructive picking of fresh tomatoes is a delicate agronomical operation, based on comprehensive information about the plant organ, such as the location of stem, peduncle, and fruits. The matching between visual information supply and information demand from the agronomical technic is the key power to promote the picking robot from the laboratory to the field. The three-dimensional pose information, containing the location of each organ of the plant, can meet the demand of agronomical technic. It is the premise of precisely handling the cluster of fruits. In order to realize the fine tomato bunch harvesting operation in a bunch, this paper proposed a three-dimensional pose detection method for tomato bunch. The method, named Tomato Pose Method (TPM), is composed of a priori geometric model, a cascaded multi-task network, and a three-dimensional reconstruction process. Based on prior knowledge and agronomic technology, this prior geometric model comprehensively and flexibly describes the spatial location information of tomato bunch. The cascaded multi-task network is designed based on hourglass structure and transfer learning, which is suitable for bounding box and key point prediction of tomato bunches in complex environments. Finally, combining the prior geometric model and the spatial position information of each key point, the tomato bunch is reconstructed. Only a medium training dataset, containing 1800 RGBD images covering changing lighting, occlusion, and various poses, is needed for training. Its success rate of TPM on two-dimensional keypoint detection is 94.02%, the accuracy of 85.77% predicted points are at medium level. And 70.05% tomato bunch with multi-pose can be constructed. More importantly, this method only needs one RGBD image taken by a commercial camera to realize the three-dimensional reconstruction of a single-bunch scenario in 1.0 s, and a multi-bunch scenario in 2.0 s. It provides comprehensive information, and provides data basis for target positioning and path planning of picking robot, which makes the non-destructive harvesting possible.
Why it matches plant phenotyping methodsトマト果房の茎・花柄・果実の三次元位置と姿勢をRGB-D画像から抽出・再構成する手法を開発し、検出精度と再構成性能を評価している。収穫ロボット応用だが、再利用可能な植物器官形態の計測が中心である。
abstractThe three-dimensional pose information, containing the location of each organ of the plant, can meet the demand of agronomical technic.
Autonomous crop monitoring is a difficult task due to the complex structure of plants. Occlusions from leaves can make it impossible to obtain complete views about all fruits of, e.g., pepper plants. Therefore, accurately estimating the shape and volume of fruits from partial information is crucial to enable further advanced automation tasks such as yield estimation and automated fruit picking. In this paper, we present an approach for mapping fruits on plants and estimating their shape by matching superellipsoids. Our system segments fruits in images and uses their masks to generate point clouds of the fruits. To combine sequences of acquired point clouds, we utilize a real-time 3D mapping framework and build up a fruit map based on truncated signed distance fields. We cluster fruits from this map and use optimized superellipsoids for matching to obtain accurate shape estimates. In our experiments, we show in various simulated scenarios with a robotic arm equipped with an RGB-D camera that our approach can accurately estimate fruit volumes. Additionally, we provide qualitative results of estimated fruit shapes from data recorded in a commercial glasshouse environment.
Why it matches plant phenotyping methods果実の画像・RGB-Dデータから形状と体積を推定する手法が研究の中心であり、植物表現型の取得・抽出に直接該当する。
abstractestimating the shape and volume of fruits from partial information is crucial
Canopy spectral reflectance can indicate both crop nutrient and canopy structural information. Differences in canopy structure can affect spectral reflectance. However, a non-imaging spectrometer cannot distinguish such differences while monitoring crop nutrients, because the results are likely to be influenced by the canopy structure. In addition, nitrogen application rate is one of the main factors influencing the canopy structure of crops. Strong correlations exist between indices of canopy structure and leaf nitrogen, and thus, these can be used to compensate for the spectral monitoring of nitrogen content in wheat leaves. In this study, canopy structural indices (CSI) such as wheat coverage, height, and textural features were obtained based on the RGB and height images obtained by the RGB-D camera. Moreover, canopy spectral reflectance was obtained by an ASD hyperspectral spectrometer, based on which two vegetation indices—ratio vegetation index (RVI) and angular insensitivity vegetation index (AIVI)—were constructed. With the vegetation indices and CSIs as input parameters, a model was established to predict the leaf nitrogen content (LNC) and leaf nitrogen accumulation (LNA) of wheat based on partial least squares (PLS) and random forest (RF) regression algorithms. The results showed that the RF model with RVI and CSI as inputs had the highest prediction accuracy for LNA, the coefficient of determination (R2) reached 0.79, and the root mean square error (RMSE) was 1.54 g/m2. The vegetation indices and coverage were relatively important features in the model. In addition, the PLS model with AIVI and CSI as input parameters had the highest prediction accuracy for LNC, with an R2 of 0.78 and an RMSE of 0.35%, among the vegetation indices. In addition, parts of both the textural and height features were important. The results suggested that PLS and RF regression algorithms can effectively integrate spectral and canopy structural information, and canopy structural information effectively supplement spectral information by improving the prediction accuracy of vegetation indices for LNA and LNC.
Why it matches plant phenotyping methodsRGB-D画像とハイパースペクトル情報を統合し、作物構造指標から小麦葉の窒素含量・蓄積量を推定する技術を開発・評価しており、表現型取得・推定手法が研究の中心である。
abstractcanopy structural indices (CSI) such as wheat coverage, height, and textural features were obtained based on the RGB and height images obtained by the RGB-D camera.
Inspired by recent promising results in sim-to-real transfer in deep learning we built a realistic simulation environment combining a Robot Operating System (ROS)-compatible physics simulator (Gazebo) with Cycles, the realistic production rendering engine from Blender. The proposed simulator pipeline allows us to simulate near-realistic RGB-D images. To showcase the capabilities of the simulator pipeline we propose a case study that focuses on indoor robotic farming. We developed a solution for sweet pepper yield estimation task. Our approach to yield estimation starts with aerial robotics control and trajectory planning, combined with deep learning-based pepper detection, and a clustering approach for counting fruit. The results of this case study show that we can combine real time dynamic simulation with near realistic rendering capabilities to simulate complex robotic systems.
Why it matches plant phenotyping methods屋内農業向けにRGB-D画像シミュレーション基盤と、コショウ果実の検出・計数による収量推定手法を開発しており、植物形質取得・推定が中心的である。
abstractThe proposed simulator pipeline allows us to simulate near-realistic RGB-D images.
Robotic High-Throughput Phenotyping (HTP) technology has been a powerful tool for selecting high-quality crop varieties among large quantities of traits. Due to the advantages of multi-view observation and high accuracy, ground HTP robots have been widely studied in recent years. In this paper, we study an ultra-narrow wheeled robot equipped with RGB-D cameras for inter-row maize HTP. The challenges of the narrow operating space, intensive light changes, and messy cross-leaf interference in rows of maize crops are considered. An in situ and inter-row stem diameter measurement method for HTP robots is proposed. To this end, we first introduce the stem diameter measurement pipeline, in which a convolutional neural network is employed to detect stems, and the point cloud is analyzed to estimate the stem diameters. Second, we present a clustering strategy based on DBSCAN for extracting stem point clouds under the condition that the stem is shaded by dense leaves. Third, we present a point cloud filling strategy to fill the stem region with missing depth values due to the occlusion by other organs. Finally, we employ convex hull and plane projection of the point cloud to estimate the stem diameters. The results show that the R2 and RMSE of stem diameter measurement are up to 0.72 and 2.95 mm, demonstrating its effectiveness.
Why it matches plant phenotyping methodsRGB-Dロボットを用いたトウモロコシ茎径の高スループット測定法を開発し、点群処理と精度評価まで行っており、表現型取得手法が研究の中心である。
abstractAn in situ and inter-row stem diameter measurement method for HTP robots is proposed.
During the last decades, consumer-grade RGB-D (red green blue-depth) cameras have gained popularity for several applications in agricultural environments. Interestingly, these cameras are used for spatial mapping that can serve for robot localization and navigation. Mapping the environment for targeted robotic applications in agricultural fields is a particularly challenging task, owing to the high spatial and temporal variability, the possible unfavorable light conditions, and the unpredictable nature of these environments. The aim of the present study was to investigate the use of RGB-D cameras and unmanned ground vehicle (UGV) for autonomously mapping the environment of commercial orchards as well as providing information about the tree height and canopy volume. The results from the ground-based mapping system were compared with the three-dimensional (3D) orthomosaics acquired by an unmanned aerial vehicle (UAV). Overall, both sensing methods led to similar height measurements, while the tree volume was more accurately calculated by RGB-D cameras, as the 3D point cloud captured by the ground system was far more detailed. Finally, fusion of the two datasets provided the most precise representation of the trees.
Why it matches plant phenotyping methodsRGB-Dカメラ、UGV、UAVによる3Dマッピング手法を比較・融合し、樹高と樹冠体積という植物形質を推定しているため、フェノタイピング手法が中心的です。
abstractThe aim of the present study was to investigate the use of RGB-D cameras and unmanned ground vehicle (UGV) for autonomously mapping the environment of commercial orchards as well as providing information about the tree height and canopy volume.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 8 Sept 2026
The use of high-throughput phenotyping with imaging and machine learning to monitor seedling growth is a tough yet intriguing subject in plant research. This has been recently addressed with low-cost RGB imaging sensors and deep learning during day time. RGB-Depth imaging devices are also accessible at low-cost and this opens opportunities to extend the monitoring of seedling during days and nights. In this article, we investigate the added value to fuse RGB imaging with depth imaging for this task of seedling growth stage monitoring. We propose a deep learning architecture along with RGB-Depth fusion to categorize the three first stages of seedling growth. Results show an average performance improvement of 5% correct recognition rate by comparison with the sole use of RGB images during the day. The best performances are obtained with the early fusion of RGB and Depth. Also, Depth is shown to enable the detection of growth stage in the absence of the light.
Why it matches plant phenotyping methodsRGB-D画像融合と深層学習による苗の生育段階推定手法を提案・比較評価しており、植物表現型取得が中心です。
abstractwe investigate the added value to fuse RGB imaging with depth imaging for this task of seedling growth stage monitoring.
Abstract Background: The truss on tomato plants is a group or cluster of smaller stems where flowers and fruit develop, while a growing truss is the most extended part of the stem. Because the state of the growing truss reacts sensitively to the surrounding environment, it is essential to control the growth in the early stages. With the recent development of IT and artificial intelligence technology in agriculture, a previous study developed a real-time acquisition and evaluation method for images using robots. Further, we used image processing to locate the growing truss and flowering rooms to extract growth information such as the height of the flower room and hard crab. Among the different vision algorithms, the CycleGAN algorithm was used to generate and transform unpaired images using generatively learning images. In this study, we developed a robot-based system for simultaneously acquiring RGB and depth images of the tomato growing truss and flower room groups. Results: The segmentation performance for approximately 35 samples was compared through the false negative (FN) and false positive (FP) indicators. For the depth camera image, we obtained FN as 17.55±3.01% and FP as 17.76±3.55%. Similarly, for CycleGAN, we obtained FN as approximately 19.24±1.45% and FP as 18.24±1.54%. As a result of image processing through depth image, IoU was 63.56 ± 8.44%, and when segmentation was performed through CycelGAN, IoU was 69.25 ± 4.42%, indicating that CycleGAN is advantageous in extracting the desired growing truss. Conclusions: The scannability was confirmed when the image scanning robot drove in a straight line through the plantation in the tomato greenhouse, which confirmed the on-site possibility of the image extraction technique using CycleGAN. In the future, the proposed approach is expected to be used in vision technology to scan the tomato growth indicators in greenhouses using an unmanned robot platform.
Why it matches plant phenotyping methodsトマトの生長トラスや花房をロボットで撮像・セグメンテーションし、生長指標を抽出する画像ベースのフェノタイピング手法とプラットフォーム開発が中心である。
abstractwe developed a robot-based system for simultaneously acquiring RGB and depth images of the tomato growing truss and flower room groups.
Rapid and accurate plant growth and biomass estimation is essential for formulating and implementing targeted forest cultivation measures. In this study, RGB-D imaging technology was used to obtain the RGB and depth imaging data for a Toona sinensis seedling canopy to estimate plant growth and aboveground biomass (AGB). Three hundred T. sinensis seedlings from 20 varieties were planted under five different drought stress treatments. The U-Net model was applied first to achieve highly accurate segmentation of plants from complex backgrounds. Simple linear regression (SLR) was used for plant height prediction, and the other three models, including multivariate linear (ML), random forest (RF) and multilayer perceptron (MLP) regression, were applied to predict the AGB and compared for optimal model selection. The results showed that the SLR model yields promising and reliable results for the prediction of plant height, with R2 and RMSE values of 0.72 and 1.89 cm, respectively. All three regression methods perform well in the prediction of AGB estimation. MLP yields the highest accuracy in predicting dry and fresh aboveground biomass compared to the other two regression models, with R2 values of 0.77 and 0.83, respectively. The combination of Gray, Green minus red (GMR) and Excess green index (ExG) was identified as the key predictor by RReliefF for predicting dry AGB. GMR was the most important in predicting fresh AGB. This study demonstrated that the merits of RGB-D and machine learning models are effective phenotyping techniques for plant height and AGB prediction, and can be used to assist dynamic responses to drought stress for breeding selection.
Why it matches plant phenotyping methodsRGB-D画像、植物セグメンテーション、回帰モデルを組み合わせ、植物高と地上部バイオマスを推定する手法が研究の中心であり、植物表現型取得・推定法として明示されています。
abstractRGB-D imaging technology was used to obtain the RGB and depth imaging data for a Toona sinensis seedling canopy to estimate plant growth and aboveground biomass (AGB).
Global warming and water resource shortage greatly influence the crop growth and negatively affect the crop yield. Breeding water–stress resistant crop varieties is one of the effective ways to handle this problem. The surface temperatures of the crop are essential for assessing their water stress resistance. Thermal imaging has been widely used to acquire the crop surface temperatures. However, most studies focus on 2D measurement. A 3D thermal imaging system is designed, and a method is developed by using the thermal and RGB-D data to acquire 3D thermal information of maize under water stress conditions. First, thermal and RGB-D cameras were used to collect the thermal and color images, and depth data of maize at the jointing stage and the thermal and color images were processed to extract the edge images of maize. Second, the KAZE feature was selected to register the edge images. Testing results showed that the KAZE feature has better performance than the SURF and BRISK features in the thermal and color image registration of maize. On the basis of the thermal and color image registration, the depth data of maize were assigned with temperature values. Third, the depth data of maize were further processed with denoising, maize extraction, amplification, smoothing, and temperature correction steps to improve the data qualities. Finally, the crop water stress index and canopy–air temperature difference values of each point were calculated. The results demonstrated that the system and the proposed method can effectively detect the water stress characteristics of maize in 3D, which can be combined with the morphological traits of leaves to synthetically analyze the water stress resistance of the crop.
Why it matches plant phenotyping methodsトウモロコシの3D熱情報を取得し、水ストレス指標を算出する画像・センサ計測法の開発が中心であり、植物表現型測定法に該当する。
abstractA 3D thermal imaging system is designed, and a method is developed by using the thermal and RGB-D data to acquire 3D thermal information of maize under water stress conditions.
We present several methods for improving plant reconstruction from multiple 3D observations. Producing 3D data useful for plant phenotyping requires proximal sensing (e.g. line scanner, depth camera) at multiple incident angles (φ) and often with multiple passes. These resulting individual point clouds must then be assembled into a single point cloud for analysis. Our interest in improving the registration of individual plants is focused specifically on observations made within field settings which present additional challenges over laboratory 3D scans, where background, overlap and light conditions can be controlled. To develop these methods, we use several season’s worth of data from the University of Arizona’s Field Scanalyzer located in Maricopa, Arizona. Our approach prioritizes: (1) plant completeness, (2) noise reduction, (3) temporal similarity and (4) computational efficiency. The first priority is accomplished simply by prioritizing individual point clouds that contain the majority of the individual plant. 3D field scanning can result in component point clouds that are from near-identical φ and cover the same portions of the individual plant. This results in both additional noise and uncertainties due to small georeferencing errors and plant movement between scans. Thus, we remove the data that is furthest in time with non-unique φ in order to achieve priorities 2 and 3. Our method results in small scene reconstruction which has low memory and computational demands. In order to improve registration further, we investigate iterative closest point (ICP) registration fitting using weights defined by crop height distributions and semantic segmentation point labeling.
Why it matches plant phenotyping methods個体植物の3Dスキャンを統合・再構成する点群登録法を開発しており、植物フェノタイピングに利用する形状データ取得・抽出が研究の中心である。
abstractWe present several methods for improving plant reconstruction from multiple 3D observations.
Beyond the use of 2D images, the analysis of 3D images is also necessary for analyzing the phenomics of crop plants. In this study, we configured a system and implemented an algorithm for the 3D image reconstruction of red pepper plant (Capsicum annuum L.), as well as its automatic analysis. A Kinect v2 with a depth sensor and a high-resolution RGB camera were used to obtain more accurate reconstructed 3D images. The reconstructed 3D images were compared with conventional reconstructed images, and the data of the reconstructed images were analyzed with respect to their directly measured features and accuracy, such as leaf number, width, and plant height. Several algorithms for image extraction and segmentation were applied for automatic analysis. The results showed that the proposed method showed an error of about 5 mm or less when reconstructing and analyzing 3D images, and was suitable for phenotypic analysis. The images and analysis algorithms obtained by the 3D reconstruction method are expected to be applied to various image processing studies.
Why it matches plant phenotyping methods植物の3D画像再構成と自動形質抽出システムの開発が中心であり、葉数・葉幅・草丈を対象に精度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明確に該当する。
abstractwe configured a system and implemented an algorithm for the 3D image reconstruction of red pepper plant (Capsicum annuum L.), as well as its automatic analysis.
Abstract Occlusion is one of the key factors affecting the success rate of vision‐based fruit‐picking robots. It is important to accurately locate and grasp the occluded fruit in field applications, However, there is yet no universal and effective solution. In this paper, a high‐precision estimation method of spatial geometric features of occluded targets based on deep learning and multisource images is presented, enabling the selective harvest robot to envision the whole target fruit as if its occlusions do not exist. First, RGB, depth and infrared images are acquired. And pixel‐level matched RGB‐D‐I fusion images are obtained by image registration. Second, aiming at the problem of detecting the occluded tomatoes in the greenhouse, an extended Mask‐RCNN network is designed to extract the target tomato. The target segmentation accuracy is improved by 7.6%. Then, for partially occluded tomatoes, a shape and position restoration method is used to recover the obscured tomato. This algorithm can extract tomato radius and centroid coordinates directly from the restored depth image. The mean Intersection over Union is 0.895, and the centroid position error is 0.62 mm for the occluded rate under 25% and the illuminance between 1 and 12 KLux. And hereby a dual‐arm robotic harvesting system is improved to achieve a picking time of 11 s per fruit, an average gripping accuracy of 8.21 mm, and an average picking success rate of 73.04%. The proposed approach realizes a high‐fidelity geometrics reconstruction instead of mere image style restoration, which endows the robot with the ability to see through obstacles in the field scenes and improves its operational success rate in its result.
Why it matches plant phenotyping methods遮蔽トマトの形状・半径・重心を画像から復元・推定する手法が中心で、単なる収穫対象の位置検出を超えた果実形態計測を技術検証している。
abstracta high‐precision estimation method of spatial geometric features of occluded targets based on deep learning and multisource images is presented
A grading method of potted Anthurium based on machine vision is proposed. A detection system is designed to acquire color images and depth images of potted Anthurium, and the three-dimensional point-cloud image is reconstructed after registration. According to the testing requirements of potted Anthurium, the minimum enclosing rectangle method is used to measure the width of crowns and spathes. The bubble sequencing method is used to measure the plant height, and the clustering segmentation method is used to calculate the number of spathes. Online automatic grading software for potted Anthurium is developed. Compared with manual measurement, the average measurement accuracies of machine vision for crown width, plant height, spathe width, and spathe number are 98.4%, 98.4%, 98.8%, and 86.7%, respectively. The accuracy rate of grading is 85.86%, which can meet the requirements of automatic grading of potted Anthurium.
Why it matches plant phenotyping methodsRGB-D画像と点群から草冠幅・草丈・仏炎苞幅・数を推定し、自動等級判定ソフトウェアを開発・検証しており、植物形質取得法が研究の中心である。
abstractA detection system is designed to acquire color images and depth images of potted Anthurium, and the three-dimensional point-cloud image is reconstructed after registration.
Obtaining 3D sensor data of complete plants or plant parts (e.g., the crop or fruit) is difficult due to their complex structure and a high degree of occlusion. However, especially for the estimation of the position and size of fruits, it is necessary to avoid occlusions as much as possible and acquire sensor information of the relevant parts. Global viewpoint planners exist that suggest a series of viewpoints to cover the regions of interest up to a certain degree, but they usually prioritize global coverage and do not emphasize the avoidance of local occlusions. On the other hand, there are approaches that aim at avoiding local occlusions, but they cannot be used in larger environments since they only reach a local maximum of coverage. In this paper, we therefore propose to combine a local, gradient-based method with global viewpoint planning to enable local occlusion avoidance while still being able to cover large areas. Our simulated experiments with a robotic arm equipped with a camera array as well as an RGB-D camera show that this combination leads to a significantly increased coverage of the regions of interest compared to just applying global coverage planning.
Why it matches plant phenotyping methods果実の位置・サイズ推定に必要な3Dセンサデータ取得を対象に、局所遮蔽回避と大域的視点計画を組み合わせる視点計画法を開発・評価しており、植物表現型取得が中心である。
abstractespecially for the estimation of the position and size of fruits, it is necessary to avoid occlusions as much as possible and acquire sensor information of the relevant parts
Reproduction assets foundThe paper's authors explicitly state that the source code of their combined local/global viewpoint planning system (used for fruit ROI coverage experiments) is publicly available on GitHub. OctoMap is a generic third-party library and is excluded.Code · publicThe source code of our system is available on GitHub 1 1
1
https://github.com/Eruvae/roi_viewpoint_planner .Open asset ↗Eruvae/roi_viewpoint_plannerlines:1-105Code / dataset availability confirmedEurope PMC · OpenAlex · checked 14 Sept 2026
The growth and the harvestability of a broccoli crop is monitored by the size of the broccoli head. This size estimation is currently done by humans, and this is inconsistent and expensive. The goal of our work was to develop a software algorithm that can estimate the size of field-grown broccoli heads based on RGB-Depth (RGB-D) images. For the algorithm to be successful, the problem of occlusion must be solved, which is the partial visibility of the broccoli head due to overlapping leaves. This partial visibility causes sizing errors. In this research, we studied the use of deep-learning algorithms to deal with occlusions. We specifically applied the Occlusion Region-based Convolutional Neural Network (ORCNN) that segmented both the visible and the amodal region of the broccoli head (which is the visible and the occluded region combined). We hypothesised that ORCNN, with its amodal segmentation, can improve the size estimation of occluded broccoli heads. The ORCNN sizing method was compared with a Mask R–CNN sizing method that only used the visible broccoli region to estimate the size. The sizing performance of both methods was evaluated on a test set of 487 broccoli images with systematic levels of leaf occlusion. With a mean sizing error of 6.4 mm, ORCNN outperformed Mask R–CNN, which had a mean sizing error of 10.7 mm. Furthermore, ORCNN had a significantly lower absolute sizing error on 161 heavily occluded broccoli heads with an occlusion rate between 50% and 90%. Our software and data set are available on https://git.wur.nl/blok012/sizecnn.
Why it matches plant phenotyping methodsRGB-D画像と深層学習によりブロッコリー頭部サイズを推定する手法を開発し、Mask R-CNNと比較検証しているため、植物形質取得法が研究の中心です。
abstractThe goal of our work was to develop a software algorithm that can estimate the size of field-grown broccoli heads based on RGB-Depth (RGB-D) images.
Kinect 3D sensing real-time acquisition algorithm that can meet the requirements of fast, accurate, and real-time acquisition of image information of crop growth laws has become the trend and necessary means of digital agricultural production management. Based on this, this paper uses Kinect real-time image generation technology to try to monitor and study the depth map of crop growth law in real time, use Kinect to obtain the algorithm of crop growth depth map, and conduct investigation and research. Real-time image acquisition research on crop growth trends provides a basis for in-depth understanding of the application of Kinect real-time image generation technology in research. Kinect image real-time acquisition algorithm is a very important information carrier in agricultural information engineering. The research results show that the real-time Kinect depth image acquisition algorithm can obtain good 3D image data information and can provide valuable data basis for the 3D reconstruction of the later crop growth model, growth status analysis, and real-time monitoring of crop diseases. The data shows that, using Kinect, the real-time feedback speed of crop growth observation can be increased by 45%, the imaging accuracy is improved by 37%, and the related operation steps are simplified by 30%. The survey results show that the crop yield can be increased by about 12%.
Why it matches plant phenotyping methodsKinectによる作物成長の深度画像をリアルタイム取得するアルゴリズムが研究の中心であり、成長状態や3D形態の表現に用いる植物フェノタイピング手法である。
abstractuse Kinect to obtain the algorithm of crop growth depth map
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · Crossref · checked 8 Sept 2026
The three-dimensional reconstruction method using RGB-D camera has a good balance in hardware cost and point cloud quality. However, due to the limitation of inherent structure and imaging principle, the acquired point cloud has problems such as a lot of noise and difficult registration. This paper proposes a 3D reconstruction method using Azure Kinect to solve these inherent problems. Shoot color images, depth images and near-infrared images of the target from six perspectives by Azure Kinect sensor with black background. Multiply the binarization result of the 8-bit infrared image with the RGB-D image alignment result provided by Microsoft corporation, which can remove ghosting and most of the background noise. A neighborhood extreme filtering method is proposed to filter out the abrupt points in the depth image, by which the floating noise point and most of the outlier noise will be removed before generating the point cloud, and then using the pass-through filter eliminate rest of the outlier noise. An improved method based on the classic iterative closest point (ICP) algorithm is presented to merge multiple-views point clouds. By continuously reducing both the size of the down-sampling grid and the distance threshold between the corresponding points, the point clouds of each view are continuously registered three times, until get the integral color point cloud. Many experiments on rapeseed plants show that the success rate of cloud registration is 92.5% and the point cloud accuracy obtained by this method is 0.789 mm, the time consuming of a integral scanning is 302 s, and with a good color restoration. Compared with a laser scanner, the proposed method has considerable reconstruction accuracy and a significantly ahead of the reconstruction speed, but the hardware cost is much lower when building a automatic scanning system. This research shows a low-cost, high-precision 3D reconstruction technology, which has the potential to be widely used for non-destructive measurement of rapeseed and other crops phenotype.
Why it matches plant phenotyping methodsRGB-D画像による植物体の3D再構成・点群登録手法を開発し、精度・成功率・処理時間を検証しており、非破壊的な作物表現型計測が中心である。
abstractThis paper proposes a 3D reconstruction method using Azure Kinect to solve these inherent problems.
Plant phenotyping relevance match · UnverifiedCrossref · Europe PMC · OpenAlex · checked 9 Sept 2026
Technology has been promoting a great transformation in farming. The introduction of robotics; the use of sensors in the field; and the advances in computer vision; allow new systems to be developed to assist processes, such as phenotyping, of crop’s life cycle monitoring. This work presents, which we believe to be the first time, a system capable of generating 3D models of non-rigid corn plants, which can be used as a tool in the phenotyping process. The system is composed by two modules: an terrestrial acquisition module and a processing module. The terrestrial acquisition module is composed by a robot, equipped with an RGB-D camera and three sets of temperature, humidity, and luminosity sensors, that collects data in the field. The processing module conducts the non-rigid 3D plants reconstruction and merges the sensor data into these models. The work presented here also shows a novel technique for background removal in depth images, as well as efficient techniques for processing these images and the sensor data. Experiments have shown that from the models generated and the data collected, plant structural measurements can be performed accurately and the plant’s environment can be mapped, allowing the plant’s health to be evaluated and providing greater crop efficiency.
Why it matches plant phenotyping methods非剛体トウモロコシの3D再構成、センサーデータ融合、背景除去、画像処理を含む表現型取得システムが研究の中心であり、植物構造形質の測定に直接利用されるため。
abstractThis work presents, which we believe to be the first time, a system capable of generating 3D models of non-rigid corn plants, which can be used as a tool in the phenotyping process.
Tea shoot detection and localization are highly challenging tasks because of varying illumination, inevitable occlusion, tiny targets, and dense growth. To achieve the automatic plucking of tea shoots in a tea garden, a reliable algorithm based on red, green, blue-depth (RGB-D) camera images was developed to detect and locate tea shoots in fields for tea harvesting robots. In this study, labeling criteria were first established for the images collected for multiple periods and varieties in the tea garden. Then, a “you only look once” (YOLO) network was used to detect tea shoot (one bud with one leaf) regions on RGB images collected by an RGB-D camera. Additionally, the detection precision for tea shoots was 93.1% and the recall rate was 89.3%. To achieve the three-dimensional (3D) localization of the plucking position, 3D point clouds of the detected target regions were acquired by fusing the depth image and RGB image captured by an RGB-D camera. Then, noise was removed using point cloud pre-processing and the point cloud of the tea shoots was obtained using Euclidean clustering processing and a target point cloud extraction algorithm. Finally, the 3D plucking position of the tea shoots was determined by combining the tea growth characteristics, point cloud features, and sleeve plucking scheme, which solved the problem that the plucking point may be invisible in fields. To verify the effectiveness of the proposed algorithm, tea shoot localization and plucking experiments were conducted in the tea garden. The plucking success rate for tea shoots was 83.18% and the average localization time for each target was about 24 ms. All the results demonstrate that the proposed method could be used for robotic tea plucking.
Why it matches plant phenotyping methodsRGB-D画像、物体検出、点群処理により茶芽の検出・3D位置推定法を開発・検証しており、植物器官の形態・位置という表現型取得が中心的です。
abstracta reliable algorithm based on red, green, blue-depth (RGB-D) camera images was developed to detect and locate tea shoots in fields for tea harvesting robots
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · Crossref · checked 14 Sept 2026
The 3D reconstruction method using RGB-D camera has a good balance in hardware cost, point cloud quality and automation. However, due to the limitation of inherent structure and imaging principle, the acquired point cloud has problems such as a lot of noise and difficult registration. This paper proposes a three-dimensional reconstruction method using Azure Kinect to solve these inherent problems. Shoot color map, depth map and near-infrared image of the target from six perspectives by Azure Kinect sensor. Multiply the 8-bit infrared image binarization with the general RGB-D image alignment result provided by Microsoft to remove ghost images and most of the background noise. In order to filter the floating point and outlier noise of the point cloud, a neighborhood maximum filtering method is proposed to filter out the abrupt points in the depth map. The floating points in the point cloud are removed before generating the point cloud, and then using the through filter filters out outlier noise. Aiming at the shortcomings of the classic ICP algorithm, an improved method is proposed. By continuously reducing the size of the down-sampling grid and the distance threshold between the corresponding points, the point clouds of each view are continuously registered three times, until get the complete color point cloud. A large number of experimental results on rape plants show that the point cloud accuracy obtained by this method is 0.739mm, a complete scan time is 338.4 seconds, and the color reduction is high. Compared with a laser scanner, the proposed method has considerable reconstruction accuracy and a significantly ahead of the reconstruction speed, but the hardware cost is much lower and it is easy to automate the scanning system. This research shows a low-cost, high-precision 3D reconstruction technology, which has the potential to be widely used for non-destructive measurement of crop phenotype.
Why it matches plant phenotyping methodsRGB-Dカメラによる植物3D再構成法を開発・検証し、作物表現型の非破壊測定への利用可能性を評価しており、表現型取得手法が研究の中心である。
abstractThis paper proposes a three-dimensional reconstruction method using Azure Kinect to solve these inherent problems.
Field / plotRGB-D / ToFFruitStem / branchObject detection2D/3D reconstructionSegmentation
In unstructured environments, harvesting robots may collide with disorderly growing branches, thus reducing the success rate of harvesting. This study introduces a fruit and branch detection and three-dimensional (3D) reconstruction method for obstacle avoidance path planning of robots. A new architecture for instance segmentation was developed by replacing the backbone of Mask R-CNN with a tiny network, referred to as “tiny Mask R-CNN”. The tiny Mask R-CNN was trained with a small number of images and used to detect guava fruits and branches. Each detected fruit and branch were converted into a 3D point cloud. It was then hypothesized that guava fruits could be represented by 3D spheres and irregular branches can be approximated by a finite number of 3D cylindrical segments. Based on the proposed hypothesis, a random sample consensus-based sphere fitting method and a principal component analysis-based cylindrical segment fitting method were investigated to reconstruct the fruits and branches from the point clouds. A guava dataset with 304 RGB-D images was collected from the fields and used to validate the developed method. The results showed that the detection F1 score of the tiny Mask R-CNN was 0.518; the F1 score for fruit reconstruction was approximately 0.851 and 0.833 under the 2D- and 3D-fruit metrics, respectively; and the F1 score for branch reconstruction was approximately 0.394 and 0.415 under the 2D- and 3D-branch metrics, respectively. These results confirm that the proposed method can effectively reconstruct the fruits and branches and can, therefore, be used to plan an obstacle avoidance path for harvesting robots.
Why it matches plant phenotyping methods果実・枝の検出と3D形状再構成という植物器官の形態計測手法を開発し、データセットで技術検証しているため、収穫ロボット用途でもフェノタイピング手法が中心である。
abstractThis study introduces a fruit and branch detection and three-dimensional (3D) reconstruction method for obstacle avoidance path planning of robots.
Plant phenotyping, that is, the quantitative assessment of plant traits including growth, morphology, physiology, and yield, is a critical aspect towards efficient and effective crop management. Currently, plant phenotyping is a manually intensive and time consuming process, which involves human operators making measurements in the field, based on visual estimates or using hand-held devices. In this work, methods for automated grapevine phenotyping are developed, aiming to canopy volume estimation and bunch detection and counting. It is demonstrated that both measurements can be effectively performed in the field using a consumer-grade depth camera mounted onboard an agricultural vehicle.
Why it matches plant phenotyping methods消費者向け深度カメラを農業車両に搭載し、圃場でブドウの樹冠体積と房の検出・計数を自動化する手法を開発しており、表現型取得が研究の中心である。
abstractIn this work, methods for automated grapevine phenotyping are developed, aiming to canopy volume estimation and bunch detection and counting.
Precision agriculture relies on the availability of accurate knowledge of crop phenotypic traits at the sub-field level. While visual inspection by human experts has been traditionally adopted for phenotyping estimations, sensors mounted on field vehicles are becoming valuable tools to increase accuracy on a narrower scale and reduce execution time and labor costs, as well. In this respect, automated processing of sensor data for accurate and reliable fruit detection and characterization is a major research challenge, especially when data consist of low-quality natural images. This paper investigates the use of deep learning frameworks for automated segmentation of grape bunches in color images from a consumer-grade RGB-D camera, placed on-board an agricultural vehicle. A comparative study, based on the estimation of two image segmentation metrics, i.e. the segmentation accuracy and the well-known Intersection over Union (IoU), is presented to estimate the performance of four pre-trained network architectures, namely the AlexNet, the GoogLeNet, the VGG16, and the VGG19. Furthermore, a novel strategy aimed at improving the segmentation of bunch pixels is proposed. It is based on an optimal threshold selection of the bunch probability maps, as an alternative to the conventional minimization of cross-entropy loss of mutually exclusive classes. Results obtained in field tests show that the proposed strategy improves the mean segmentation accuracy of the four deep neural networks in a range between 2.10 and 8.04%. Besides, the comparative study of the four networks demonstrates that the best performance is achieved by the VGG19, which reaches a mean segmentation accuracy on the bunch class of 80.58%, with IoU values for the bunch class of 45.64%.
Why it matches plant phenotyping methodsブドウ房を対象とした画像セグメンテーション手法の開発・比較・性能評価が中心であり、植物器官の形態状態を抽出するフェノタイピング手法に該当する。
abstractThis paper investigates the use of deep learning frameworks for automated segmentation of grape bunches in color images from a consumer-grade RGB-D camera
In this research, multispectral and depth images were utilized for tea moisture content detection, the problems on leaf surface orientation and detection height were studied specifically. For the leaf surface orientation issue, multispectral images (25 bands) of the front surface and back surface of tea leaves were collected. Based on the spectra with same surface orientation, regression models of tea moisture content were established. The R²P values of LSSVR models reach 0.77 and 0.68 for the front surface and back surface, respectively. To distinguish the surface orientation of tea leaves, an LDA classifier was built based on spectral band ratio information. The overall classification accuracy reaches 87.8%. The distribution map of tea moisture content was successfully generated by importing the spectra into the classifier and the regression model. For the detection height issue, the multispectral image and depth image of tea leaves were collected simultaneously. First, an experiment was designed to figure out the attenuation coefficient of each band and the calibration model of detection height. Then, the detection height information was introduced into each pixel of the multispectral image by image registration. According to the detection height and calibration model, the spectrum of each pixel was calibrated. Finally, through importing the modified spectra into the classifier and regression model, the visual detection of tea moisture content was realized with detection height calibration. This research promoted the practicability of tea moisture content detection, and improved the visualization detection technology based on the fusion of multispectral image and depth image.
Why it matches plant phenotyping methods茶葉の水分含量という植物器官の状態を、マルチスペクトル・深度画像、分類、回帰、画像登録、検出高さ補正により推定・可視化する方法が研究の中心である。
abstractmultispectral and depth images were utilized for tea moisture content detection
Abstract Canopy height is an essential trait in high throughput phenotyping that is often only captured as a single point, which is not always representative of canopy height. New 3D depth cameras such as the RealSense D415 (Intel Corporation, Santa Clara, CA, USA) may provide a fast and affordable solution for measuring height from portable, ground‐based phenomics systems. Our goal was to determine if the D415 was effective at measuring crop heights under field conditions in wheat ( Triticum aestivum ) and soybean ( Glycine max ) plots. The D415 camera was integrated into our PlotCam platform using the open software development kit from Intel. Distance arrays were captured for each plot at weekly intervals over the growing season. These were compared to canopy heights measured using a single point LiDAR (SPL) system operated by hand. Over the growing season the D415 heights were significantly correlated with the SPL heights in both wheat and soybean with coefficients of 0.77 and 0.95 and NRMSE 0.23 and 0.17 m, respectively. Early season D415 height measurements were not as similar to the SPL as the mid‐and late‐season measurements in wheat and soybean. The relatively low cost and open software development kit of the D415 makes it a promising tool for high throughput phenotyping applications.
Why it matches plant phenotyping methods低コスト深度カメラを用いた作物キャノピー高測定法を開発・検証し、LiDARと比較評価しているため、表現型取得手法が研究の中心です。
abstractThese were compared to canopy heights measured using a single point LiDAR (SPL) system operated by hand.
Spraying in farmlands and greenhouses is not only labor and time consuming, but also it has adverse effects on human health. The best solution is to use mechanized devices and robots to automate the spraying operation. In this study, a robot with four degrees of freedom performs spraying through calculating the plant bulk volume. The robot travels along the plant rows and stops when it detects plant stems and opens its manipulator step by step so that the Kinect v. 1 camera can capture deep RGB¹1- Red, Green, Blue. colored images from the various sections. After calculating the bulk volume of the plant, it starts spraying. The results showed that the robot is capable of calculating the volume efficiently. The average operation time of the robot detecting plant, opening manipulator, imaging, estimating volume, and spraying liquid according to the estimated volume is 54 s for a plant of 1.7 m height. The detection error of the robot is estimated to be 19%, which is negligible. Assessment of liquid-sensitive papers located around the plants shows that the robot’s liquid-spraying quality is not reduced at high levels of the plant. The robot can detect the plant’s height and preserve its desired spraying quality in its total height.
Why it matches plant phenotyping methodsKinect画像から植物の体積・高さを推定し、その推定値に基づき散布量を制御するロボット手法が研究の中心であり、植物形態形質の取得と技術性能を評価している。
abstracta robot with four degrees of freedom performs spraying through calculating the plant bulk volume
The accurate estimation of nitrogen accumulation is of great significance to nitrogen fertilizer management in wheat production. To overcome the shortcomings of spectral technology, which ignores the anisotropy of canopy structure when predicting the nitrogen accumulation in wheat, resulting in low accuracy and unstable prediction results, we propose a method for predicting wheat nitrogen accumulation based on the fusion of spectral and canopy structure features. After depth images are repaired using a hole-filling algorithm, RGB images and depth images are fused through IHS transformation, and textural features of the fused images are then extracted in order to express the three-dimensional structural information of the canopy. The fused images contain depth information of the canopy, which breaks through the limitation of extracting canopy structure features from a two-dimensional image. By comparing the experimental results of multiple regression analyses and BP neural networks, we found that the characteristics of the canopy structure effectively compensated for the model prediction of nitrogen accumulation based only on spectral characteristics. Our prediction model displayed better accuracy and stability, with prediction accuracy values (R2) based on BP neural network for the leaf layer nitrogen accumulation (LNA) and shoot nitrogen accumulation (SNA) during a full growth period of 0.74 and 0.73, respectively, and corresponding relative root mean square errors (RRMSEs) of 40.13% and 35.73%.
Why it matches plant phenotyping methodsRGB-D画像とスペクトル画像の融合、特徴抽出、予測モデルによりコムギの窒素蓄積を推定する方法が研究の中心であり、植物形質の取得・推定手法に該当する。
abstractwe propose a method for predicting wheat nitrogen accumulation based on the fusion of spectral and canopy structure features.
The use of 3D sensors combined with appropriate data processing and analysis has provided tools to optimise agricultural management through the application of precision agriculture. The recent development of low-cost RGB-Depth cameras has presented an opportunity to introduce 3D sensors into the agricultural community. However, due to the sensitivity of these sensors to highly illuminated environments, it is necessary to know under which conditions RGB-D sensors are capable of operating. This work presents a methodology to evaluate the performance of RGB-D sensors under different lighting and distance conditions, considering both geometrical and spectral (colour and NIR) features. The methodology was applied to evaluate the performance of the Microsoft Kinect v2 sensor in an apple orchard. The results show that sensor resolution and precision decreased significantly under middle to high ambient illuminance (>2000 lx). However, this effect was minimised when measurements were conducted closer to the target. In contrast, illuminance levels below 50 lx affected the quality of colour data and may require the use of artificial lighting. The methodology was useful for characterizing sensor performance throughout the full range of ambient conditions in commercial orchards. Although Kinect v2 was originally developed for indoor conditions, it performed well under a range of outdoor conditions.
Why it matches plant phenotyping methodsRGB-Dセンサーの性能を、果樹キャノピーの3D形状・色・NIR特徴の取得という植物フェノタイピング用途で、照明・距離条件下で評価する方法論が中心である。
abstractThis work presents a methodology to evaluate the performance of RGB-D sensors under different lighting and distance conditions, considering both geometrical and spectral (colour and NIR) features.
Reproduction assets foundThe paper's Kinect Evaluation in Orchard conditions (KEvOr) dataset of RGB/NIR/point-cloud captures from an apple orchard is publicly deposited on Zenodo, and the authors' MATLAB analysis code for the sensor evaluation is publicly available on GitHub. Both are paper-specific, public, and actionable.Code · publicA MATLAB® (R2020a, Math Works Inc., Natick, MA, USA) code was developed to analyse all
the data and provide the sensor evaluation results. This code has been made publicly available at
https://github.com/GRAP-UdL-AT/RGBD_sensors_evaluation_in_Orchards [35].Open asset ↗GitHub · GRAP-UdL-AT/RGBD_sensors_evaluation_in_Orchardspdf-page:6 lines:1-60Plant phenotyping relevance match · UnverifiedOpenAlex · checked 13 Sept 2026
A non-destructive measuring technique was applied to test major vine geometric traits on measurements collected by a contactless sensor. Three-dimensional optical sensors have evolved over the past decade, and these advancements may be useful in improving phenomics technologies for other crops, such as woody perennials. Red, green and blue-depth (RGB-D) cameras, namely Microsoft Kinect, have a significant influence on recent computer vision and robotics research. In this experiment an adaptable mobile platform was used for the acquisition of depth images for the non-destructive assessment of branch volume (pruning weight) and related to grape yield in vineyard crops. Vineyard yield prediction provides useful insights about the anticipated yield to the winegrower, guiding strategic decisions to accomplish optimal quantity and efficiency, and supporting the winegrower with decision-making. A Kinect v2 system on-board to an on-ground electric vehicle was capable of producing precise 3D point clouds of vine rows under six different management cropping systems. The generated models demonstrated strong consistency between 3D images and vine structures from the actual physical parameters when average values were calculated. Correlations of Kinect branch volume with pruning weight (dry biomass) resulted in high coefficients of determination (R2 = 0.80). In the study of vineyard yield correlations, the measured volume was found to have a good power law relationship (R2 = 0.87). However due to low capability of most depth cameras to properly build 3-D shapes of small details the results for each treatment when calculated separately were not consistent. Nonetheless, Kinect v2 has a tremendous potential as a 3D sensor in agricultural applications for proximal sensing operations, benefiting from its high frame rate, low price in comparison with other depth cameras, and high robustness.
Why it matches plant phenotyping methodsRGB-Dセンサー搭載移動プラットフォームでブドウ樹の枝体積を非破壊推定し、剪定重量および収量との相関で技術性能を検証しており、植物形質取得法が研究の中心である。
abstractA non-destructive measuring technique was applied to test major vine geometric traits on measurements collected by a contactless sensor.
Accurately detecting and counting fruits during plant growth using imaging and computer vision is of importance not only from the point of view of reducing labor intensive manual measurements of phenotypic information, but also because it is a critical step toward automating processes such as harvesting. Deep learning based methods have emerged as the state-of-the-art techniques in many problems in image segmentation and classification, and have a lot of promise in challenging domains such as agriculture, where they can deal with the large variability in data better than classical computer vision methods. This paper reports results on the detection of tomatoes in images taken in a greenhouse, using the MaskRCNN algorithm, which detects objects and also the pixels corresponding to each object. Our experimental results on the detection of tomatoes from images taken in greenhouses using a RealSense camera are comparable to or better than the metrics reported by earlier work, even though those were obtained in laboratory conditions or using higher resolution images. Our results also show that MaskRCNN can implicitly learn object depth, which is necessary for background elimination.
Why it matches plant phenotyping methods温室画像からトマト果実を検出・計数する画像解析手法を開発・評価しており、果実数という植物形質の取得が中心的な技術課題であるため。
abstractAccurately detecting and counting fruits during plant growth using imaging and computer vision is of importance not only from the point of view of reducing labor intensive manual measurements of phenotypic information
RGB-D cameras have been successfully used for indoor High-ThroughpuT Phenotyping (HTTP). However, their capability and feasibility for in-field HTTP still need to be evaluated, due to the noise and disturbances generated by unstable illumination, specular reflection, and diffuse reflection, etc. To solve these problems, we evaluated the depth-ranging performances of two consumer-level RGB-D cameras (RealSense D435i and Kinect V2) under in-field HTTP scenarios, and proposed a strategy to compensate the depth measurement error. For performance evaluation, we focused on determining their optimal ranging areas for different crop organs. Based on the evaluation results, we proposed a brightness-and-distance-based Support Vector Regression Strategy, to compensate the ranging error. Furthermore, we analyzed the depth filling rate of two RGB-D cameras under different lighting intensities. Experimental results showed that: 1) For RealSense D435i, its effective ranging area is [0.160, 1.400] m, and in-field filling rate is approximately 90%. 2) For Kinect V2, it has a high ranging accuracy in the [0.497, 1.200] m, but its in-field filling rate is less than 24.9%. 3) Our error compensation model can effectively reduce the influences of lighting intensity and target distance. The maximum MSE and minimum R2 of this model are 0.029 and 0.867, respectively. To sum up, RealSense D435i has better ranging performances than Kinect V2 on in-field HTTP.
Why it matches plant phenotyping methods圃場ハイスループット表現型解析ロボット向けRGB-Dカメラの性能評価と深度誤差補償法の開発が研究の中心であり、植物器官の距離計測という表現型取得に直接関係する。
abstractwe evaluated the depth-ranging performances of two consumer-level RGB-D cameras (RealSense D435i and Kinect V2) under in-field HTTP scenarios, and proposed a strategy to compensate the depth measurement error.
Effective segmentation of plant leaves is very necessary for non-contact extraction of plant leaf phenotype, especially leaf phenotype under environmental stress. However, the phenotype of leaves will change due to the influence of the environment, which increases the difficulty of detection. In this study, we proposed an accurate automatic segmentation method that combines Mask R-CNN with Density-Based Spatial Clustering of Applications with Noise (DBSCAN) clustering algorithm based on RGB-D camera to segment overlapped poplar seedling leaves under heavy metal stress. Firstly, an effective encoding method of depth information was used to facilitate the feature extraction of depth information. Next, we deployed Mask R-CNN to train the RGB-D data and fuse their features in the FPN structure to obtain more accurate leaf areas. Based on the detected leaf areas and depth data, DBSCAN based on manifold distance was then applied to segment a single leaves from overlapping leaves in the detected areas. Several analyses were performed to evaluate the performance of the proposed method, including the comparison of our network with classic Mask R-CNN and the comparison of DBSCAN based on manifold distance with other classic clustering methods. We used the pixel-wise Intersection over Union (p-IoU) to evaluate the detection results more accurately. In the experiments, the obtained p-IoU of normal and stressed leaves was 0.885 and 0.874, respectively, with corresponding mean accuracy values of 0.897 and 0.888. From our experimental results, it can be concluded that the proposed method can automatically detect leaves with high accuracy, which can be applied to 3-D leaf phenotype research and automatic plant de-leafing.
Why it matches plant phenotyping methodsRGB-D画像、Mask R-CNN、DBSCANを組み合わせ、重なったポプラ葉のセグメンテーション手法を開発・評価しており、葉形質抽出が中心である。
abstractwe proposed an accurate automatic segmentation method that combines Mask R-CNN with Density-Based Spatial Clustering of Applications with Noise (DBSCAN) clustering algorithm based on RGB-D camera to segment overlapped poplar seedling leaves under heavy metal stress.
Modern agricultural applications require knowledge about the position and size of fruits on plants. However, occlusions from leaves typically make obtaining this information difficult. We present a novel viewpoint planning approach that builds up an octree of plants with labeled regions of interest (ROIs), i.e., fruits. Our method uses this octree to sample viewpoint candidates that increase the information around the fruit regions and evaluates them using a heuristic utility function that takes into account the expected information gain. Our system automatically switches between ROI targeted sampling and exploration sampling, which considers general frontier voxels, depending on the estimated utility. When the plants have been sufficiently covered with the RGB-D sensor, our system clusters the ROI voxels and estimates the position and size of the detected fruits. We evaluated our approach in simulated scenarios and compared the resulting fruit estimations with the ground truth. The results demonstrate that our combined approach outperforms a sampling method that does not explicitly consider the ROIs to generate viewpoints in terms of the number of discovered ROI cells. Furthermore, we show the real-world applicability by testing our framework on a robotic arm equipped with an RGB-D camera installed on an automated pipe-rail trolley in a capsicum glasshouse.
Why it matches plant phenotyping methods果実の位置・サイズという植物器官形質をRGB-Dセンサで取得・推定する視点計画法を開発し、シミュレーションと実環境で検証しており、フェノタイピング手法が中心である。
abstractWe present a novel viewpoint planning approach that builds up an octree of plants with labeled regions of interest (ROIs), i.e., fruits.
Reproduction assets foundThe paper's viewpoint-planning system source code and the simulated capsicum plant environments used in the experiments are publicly available on GitHub. OctoMap is a generic third-party library, not a paper-specific asset.Code · publicThe source code of our system is available on GitHub 1 1
1
https://github.com/Eruvae/roi_viewpoint_planner .Open asset ↗Eruvae/roi_viewpoint_plannerlines:73-113Plant phenotyping relevance match · UnverifiedEurope PMC · OpenAlex · checked 15 Sept 2026
Field / plotRGB-D / ToFFruitCountingObject detection
The accurate and reliable fruit detection in orchards is one of the most crucial tasks for supporting higher level agriculture tasks such as yield mapping and robotic harvesting. However, detecting and counting small fruit is a very challenging task under variable lighting conditions, low-resolutions and heavy occlusion by neighboring fruits or foliage. To robustly detect small fruits, an improved method is proposed based on multiple scale faster region-based convolutional neural networks (MS-FRCNN) approach using the color and depth images acquired with an RGB-D camera. The architecture of MS-FRCNN is improved to detect lower-level features by incorporating feature maps from shallower convolution feature maps for regions of interest (ROI) pooling. The detection framework consists of three phases. Firstly, multiple scale feature extractors are used to extract low and high features from RGB and depth images respectively. Then, RGB-detector and depth-detector are trained separately using MS-FRCNN. Finally, late fusion methods are explored for combining the RGB and depth detector. The detection framework was demonstrated and evaluated on two datasets that include passion fruit images under variable illumination conditions and occlusion. Compared with the faster R-CNN detector of RGB-D images, the recall, the precision and F1-score of MS-FRCNN method increased from 0.922 to 0.962, 0.850 to 0.931 and 0.885 to 0.946, respectively. Furthermore, the MS-FRCNN method effectively improves small passion fruit detection by achieving 0.909 of the F1 score. It is concluded that the detector based on MS-FRCNN can be applied practically in the actual orchard environment.
Why it matches plant phenotyping methodsRGB-D画像から果実の検出・計数という植物器官の形態・収量関連形質を抽出する手法を開発・評価しており、植物フェノタイピング手法が中心である。
abstractTo robustly detect small fruits, an improved method is proposed based on multiple scale faster region-based convolutional neural networks (MS-FRCNN) approach using the color and depth images acquired with an RGB-D camera.
Fruit detection and localization are essential for future agronomic management of fruit crops such as yield prediction, yield mapping and automated harvesting. However, to perform robust and efficient fruit detection and localization in orchard is a challenging task under variable illumination, low-resolutions and heavy occlusion by neighboring fruits, foliage, or branches. Therefore, researches of fruit detection and localization by getting more information of objects are essential. RGB-D (Red, Green, Blue -Depth) cameras are promising sensors and widely used in fruit detection and localization given that they provide depth information and infrared information in addition to RGB information. After presenting a discussion on the advantages and disadvantages of RGB-D cameras with different depth measurement principles and application fields, this paper reviews various types of RGB-D sensor systems and image processing methods used for fruit detection and localization in the field. Finally, major challenges for the successful application of RGB-D camera-based machine vision system, and potential future directions for the research and development in this area are discussed.
Why it matches plant phenotyping methodsRGB-Dセンサーと画像処理による果実の検出・位置推定手法を中心にレビューしており、植物器官の観測・抽出法が主題である。
abstractthis paper reviews various types of RGB-D sensor systems and image processing methods used for fruit detection and localization in the field.
To help address the global growing demand for food and fiber, selective breeding programs aim to cultivate crops with higher yields and more resistance to stress. Measuring phenotypic traits needed for breeding programs is usually done manually and is labor-intensive, subjective, and lacks adequate temporal resolution. This paper presents a Multipurpose Autonomous Robot of Intelligent Agriculture (MARIA), an open source differential drive robot that is able to navigate autonomously indoors and outdoors while conducting plant morphological trait phenotyping and soil sensing. For the design of the rover, a drive system was developed using the Robot Operating System (ROS), which allows for autonomous navigation using Global Navigation Satellite Systems (GNSS). For phenotyping, the robot was fitted with an actuated LiDAR unit and a depth camera that can estimate morphological traits of plants such as volume and height. A three degree-of-freedom manipulator mounted on the mobile platform was designed using Dynamixel servos that can perform soil sensing and sampling using off-the-shelf and 3D printed components. MARIA was able to navigate both indoors and outdoors with an RMSE of 0.0156 m and 0.2692 m, respectively. Additionally, the onboard actuated LiDAR sensor was able to estimate plant volume and height with an average error of 1.76% and 3.2%, respectively. The manipulator performance tests on soil sensing was also satisfactory. This paper presents a design for a differential drive mobile robot built from off-the-shelf components that makes it replicable and available for implementation by other researchers. The validation of this system suggests that it may be a valuable solution to address the phenotyping bottleneck by providing a system capable of navigating through crop rows or a greenhouse while conducting phenotyping and soil measurements.
Why it matches plant phenotyping methods植物形態形質の取得を中核とする自律移動ロボットを開発し、LiDAR・深度カメラによる草高・体積推定を検証しているため、植物フェノタイピング手法・プラットフォーム研究に該当する。
abstractThis paper presents a Multipurpose Autonomous Robot of Intelligent Agriculture (MARIA), an open source differential drive robot that is able to navigate autonomously indoors and outdoors while conducting plant morphological trait phenotyping and soil sensing.
Frequent measurements of the plant phenotypes make it possible to monitor plant status during the growing season. Stem diameter is an important proxy for overall plant biomass and health. However, the manual measurement of stem diameter in plants is time consuming, error prone, and laborious. The use of agricultural robots to automatically collect plant phenotypic data for trait measurements can overcome many of the drawbacks of manual phenotyping. The objective of this research was to develop a robotic system that can automatically detect and grasp the stem, and measure its diameter of maize and sorghum plants. The robotic system comprises of a four degree of freedom robotic manipulator, a time-of-flight camera for vision system, and a linear potentiometer sensor to measure the stem diameter. Deep learning and conventional image processing were used to detect stem in images and find grasping point of stem, respectively. An experiment was conducted in a greenhouse using maize and sorghum plants to evaluate the performance of the robotic system. The system demonstrated successful grasping of stem and a high correlation between manual and robotic measurements of diameter depicting its ability to be used as a prototype to integrate other sensors to measure different physiological and chemical attributes of the stem.
Why it matches plant phenotyping methodsロボットによる茎径という植物形質の自動取得システムを開発し、手動測定との性能比較で評価しており、フェノタイピング手法が研究の中心です。
abstractThe objective of this research was to develop a robotic system that can automatically detect and grasp the stem, and measure its diameter of maize and sorghum plants.
Plant phenotyping relevance match · UnverifiedOpenAlex · Europe PMC · checked 15 Sept 2026
Plant phenotyping is the task of measuring plant attributes mainly for agricultural purposes. We term length phenotyping the task of measuring the length of a plant part of interest. The recent rise of low cost RGB-D sensors and accurate deep artificial neural networks provides new opportunities for length phenotyping. We present a general technique for length phenotyping based on three stages: object detection, point of interest identification, and a 3D measurement phase. We address object detection and interest point identification by training network models for each task, and develop a robust de-projection procedure for the 3D measurement stage. We apply our method to three real world tasks: measuring the height of a banana tree, the length and width of banana leaves in potted plants, and the length of cucumbers fruits in field conditions. The three tasks were solved using the same pipeline with minor adaptations, indicating the method’s general potential. The method is stagewise analyzed and shown to be preferable to alternative algorithms, obtaining error of less than 10% deviation in all tasks. For leaves’ length and width, the measurements are shown to be useful for further phenotyping of plant treatment and mutant classification.
Why it matches plant phenotyping methods植物部位の長さをRGB-Dセンサー、物体検出、関心点検出、3D計測で推定する汎用フェノタイピング手法の開発・評価が中心である。
abstractWe present a general technique for length phenotyping based on three stages: object detection, point of interest identification, and a 3D measurement phase.
The effect of camera viewpoint and fruit orientation on the performance of a sweet pepper maturity level classification algorithm was evaluated. Image datasets of sweet peppers harvested from a commercial greenhouse were collected using two different methods, resulting in 789 RGB-Red Green Blue (images acquired in a photocell) and 417 RGB-D-Red Green Blue-Depth (images acquired by a robotic arm in the laboratory), which are published as part of this paper. Maturity level classification was performed using a random forest algorithm. Classifications of maturity level from different camera viewpoints, using a combination of viewpoints, and different fruit orientations on the plant were evaluated and compared to manual classification. Results revealed that: (1) the bottom viewpoint is the best single viewpoint for maturity level classification accuracy; (2) information from two viewpoints increases the classification by 25 and 15 percent compared to a single viewpoint for red and yellow peppers, respectively, and (3) classification performance is highly dependent on the fruit's orientation on the plant.
Why it matches plant phenotyping methodsスイートペッパー果実の成熟度という植物形質を、画像取得視点・果実向き・分類アルゴリズムの性能評価により推定しており、フェノタイピング手法の評価が中心である。
abstractThe effect of camera viewpoint and fruit orientation on the performance of a sweet pepper maturity level classification algorithm was evaluated.
The key technology for a fruit picking robot is to identify fruits in different occlusion states. Based on the mask regional convolutional neural network (Mask R-CNN) and a branch segment merging algorithm, an integrated system was developed to simultaneously detect and measure citrus fruits and branches. A training dataset was constructed for fruit and tree appearance, including single fruit, multiple fruits, occluded fruits, branches and trunk. A segmental labeling method for random and irregular branches is proposed to improve the precision of the Mask R-CNN. Based on the segmental mask regions identified by this model, a more precise bounding box is obtained by calculating the minimum enclosing rectangle of mask regions. Then, a branch segment merging algorithm reconstructs branches and the trunk. Diameters of fruits and branches are obtained by mapping the color image onto the depth image. The average precision of fruit and branch recognition are 88.15% and 96.27%, respectively. The average measurement error of fruits’ transverse diameters, fruits’ longitudinal diameters, and branch diameters are 2.52, 2.29, and 1.17 mm, respectively. Experiments show the detection system has good performance for all types of fruits and occlusions. This vision system can effectively help the robot to plan the appropriate picking path and avoid obstacles.
Why it matches plant phenotyping methodsMask R-CNNと深度画像を用いて果実・枝を検出し、直径を定量する画像ベースの植物形質計測システムが中心であり、単なる収穫対象の位置検出を超えている。
abstractan integrated system was developed to simultaneously detect and measure citrus fruits and branches.
In the last decade, three-dimensional reconstruction of plants has gained a noticeable importance, in particular for the possibility of collecting data correlated to biomass, leaves area, etc. Different sensing technologies are available for 3D imaging, as for instance laser scanning, or stereoscopic reconstruction: however, practical application is still limited by high costs, or high speed data processing demand.For the present work depth sensing cameras technology is implemented.Measurements were repeated on 17 different dates, between April and June, on a jujube (Ziziphus jujube) plant, collecting 3D scans through a Kinect I sensor.3D images were analysed in order to estimate the three-dimensional volume (Vk) of the canopy and the leaf area, and the results were compared with biomass related data arising from hand measurements (volume and leaf area index, LAI).
Why it matches plant phenotyping methodsKinect深度カメラによる3D植物画像取得と、樹冠体積・葉面積の推定および手測定との比較が研究の中心であり、植物表現型計測手法の適用・評価に該当する。
abstractdepth sensing cameras technology is implemented
Plant height and leaf area are important morphological properties of leafy vegetable seedlings, and they can be particularly useful for plant growth and health research. The traditional measurement scheme is time-consuming and not suitable for continuously monitoring plant growth and health. Individual vegetable seedling quick segmentation is the prerequisite for high-throughput seedling phenotype data extraction at individual seedling level. This paper proposes an efficient learning- and model-free 3D point cloud data processing pipeline to measure the plant height and leaf area of every single seedling in a plug tray. The 3D point clouds are obtained by a low-cost red–green–blue (RGB)-Depth (RGB-D) camera. Firstly, noise reduction is performed on the original point clouds through the processing of useable-area filter, depth cut-off filter, and neighbor count filter. Secondly, the surface feature histograms-based approach is used to automatically remove the complicated natural background. Then, the Voxel Cloud Connectivity Segmentation (VCCS) and Locally Convex Connected Patches (LCCP) algorithms are employed for individual vegetable seedling partition. Finally, the height and projected leaf area of respective seedlings are calculated based on segmented point clouds and validation is carried out. Critically, we also demonstrate the robustness of our method for different growth conditions and species. The experimental results show that the proposed method could be used to quickly calculate the morphological parameters of each seedling and it is practical to use this approach for high-throughput seedling phenotyping.
Why it matches plant phenotyping methodsRGB-D点群を用いて個体ごとの草丈・葉面積を抽出する高スループット表現型計測パイプラインを開発し、異なる生育条件・種で検証しており、表現型取得手法が中心である。
abstractThis paper proposes an efficient learning- and model-free 3D point cloud data processing pipeline to measure the plant height and leaf area of every single seedling in a plug tray.
With today's environmental challenges, developing sustainable energy sources is crucial. From this perspective, woody biomass has been, and continues to be, a significant research interest. The goal of this research was to develop new technology for mapping willow tree yield grown in a short-rotation forestry (SRF) system. The system gathered the physical characteristics of willow trees on-the-go, while the trees were being harvested. Features assessed include the number of trees harvested and their diameter. To complete this task, a machine-vision system featuring an RGB-D stereovision camera was built. The system tagged these data with the corresponding geographical coordinates using a Global Navigation Satellite System (GNSS) receiver. The proposed yield-mapping system showed promising detection results considering the complex background and variable light conditions encountered in the outdoors. Of the 40 randomly selected and manually observed trees in a row, 36 were successfully detected, yielding a 90% detection rate. The correctly detected tree rate of all trees within the scenes was actually 71.8% since the system tended to be sensitive to branches, thus, falsely detecting them as trees. Manual validation of the diameter estimation function showed a poor coefficient of determination and a root mean square error (RMSE) of 10.7 mm.
Why it matches plant phenotyping methodsRGB-Dステレオビジョンで収穫中のヤナギの個体数と直径を取得・推定する収量マッピング技術を開発し、検出性能と直径推定を検証しており、植物表現型取得が中心である。
abstractThe goal of this research was to develop new technology for mapping willow tree yield grown in a short-rotation forestry (SRF) system.
A prediction method of mass and volume of cherry tomato based on a computer vision system and machine learning algorithms were introduced in this study. The relation between tomato mass and volume was established as M=1.312V0.9551, and was used to estimate mass on a test dataset at an R2 of 0.9824 and RMSE of 15.84g. Depth images of tomatoes at different orientations were acquired and features extracted by image processing techniques. Five regression prediction models based on 2D and 3D image features were developed. The RBF-SVM outperformed all explored models with an accuracy of 0.9706 (only 2D features) and 0.9694 (all features) in mass and volume estimation respectively. The model predicted mass or volume can then be applied to the established mass-volume power function. This introduced system can be applied as a non-destructive, accurate and consistent technique to in-line sorting and grading of cherry tomatoes based on mass, volume or density.
Why it matches plant phenotyping methodsコンピュータビジョンと機械学習により、トマト果実の体積・質量という植物器官形質を推定する手法を開発・評価しており、表現型取得が研究の中心です。
abstractA prediction method of mass and volume of cherry tomato based on a computer vision system and machine learning algorithms were introduced in this study.
Measurement of plant nitrogen (N), phosphorus (P), and potassium (K) levels are important for determining precise fertilization management approaches for crops cultivated in greenhouses. To accurately, rapidly, stably, and nondestructively measure the NPK levels in tomato plants, a nondestructive determination method based on multispectral three-dimensional (3D) imaging was proposed. Multiview RGB-D images and multispectral images were synchronously collected, and the plant multispectral reflectance was registered to the depth coordinates according to Fourier transform principles. Based on the Kinect sensor pose estimation and self-calibration, the unified transformation of the multiview point cloud coordinate system was realized. Finally, the iterative closest point (ICP) algorithm was used for the precise registration of multiview point clouds and the reconstruction of plant multispectral 3D point cloud models. Using the normalized grayscale similarity coefficient, the degree of spectral overlap, and the Hausdorff distance set, the accuracy of the reconstructed multispectral 3D point clouds was quantitatively evaluated, the average value was 0.9116, 0.9343 and 0.41 cm, respectively. The results indicated that the multispectral reflectance could be registered to the Kinect depth coordinates accurately based on the Fourier transform principles, the reconstruction accuracy of the multispectral 3D point cloud model met the model reconstruction needs of tomato plants. Using back-propagation artificial neural network (BPANN), support vector machine regression (SVMR), and gaussian process regression (GPR) methods, determination models for the NPK contents in tomato plants based on the reflectance characteristics of plant multispectral 3D point cloud models were separately constructed. The relative error (RE) of the N content by BPANN, SVMR and GPR prediction models were 2.27%, 7.46% and 4.03%, respectively. The RE of the P content by BPANN, SVMR and GPR prediction models were 3.32%, 8.92% and 8.41%, respectively. The RE of the K content by BPANN, SVMR and GPR prediction models were 3.27%, 5.73% and 3.32%, respectively. These models provided highly efficient and accurate measurements of the NPK contents in tomato plants. The NPK contents determination performance of these models were more stable than those of single-view models.
Why it matches plant phenotyping methodsトマトのNPK含量という植物状態を、マルチスペクトル・3D画像、点群再構成、画像位置合わせ、回帰モデルで非破壊推定する方法を開発・定量評価しており、フェノタイピング手法が中心である。
abstracta nondestructive determination method based on multispectral three-dimensional (3D) imaging was proposed
BACKGROUND: Assessing the impact of the environment on plant performance requires growing plants under controlled environmental conditions. Plant phenotypes are a product of genotype × environment (G × E), and the Enviratron at Iowa State University is a facility for testing under controlled conditions the effects of the environment on plant growth and development. Crop plants (including maize) can be grown to maturity in the Enviratron, and the performance of plants under different environmental conditions can be monitored 24 h per day, 7 days per week throughout the growth cycle. RESULTS: The Enviratron is an array of custom-designed plant growth chambers that simulate different environmental conditions coupled with precise sensor-based phenotypic measurements carried out by a robotic rover. The rover has workflow instructions to periodically visit plants growing in the different chambers where it measures various growth and physiological parameters. The rover consists of an unmanned ground vehicle, an industrial robotic arm and an array of sensors including RGB, visible and near infrared (VNIR) hyperspectral, thermal, and time-of-flight (ToF) cameras, laser profilometer and pulse-amplitude modulated (PAM) fluorometer. The sensors are autonomously positioned for detecting leaves in the plant canopy, collecting various physiological measurements based on computer vision algorithms and planning motion via "eye-in-hand" movement control of the robotic arm. In particular, the automated leaf probing function that allows the precise placement of sensor probes on leaf surfaces presents a unique advantage of the Enviratron system over other types of plant phenotyping systems. CONCLUSIONS: The Enviratron offers a new level of control over plant growth parameters and optimizes positioning and timing of sensor-based phenotypic measurements. Plant phenotypes in the Enviratron are measured in situ-in that the rover takes sensors to the plants rather than moving plants to the sensors.
Why it matches plant phenotyping methodsロボット rover、複数センサー、コンピュータビジョン、葉面プロービングを統合した植物フェノタイピング基盤の開発・実証が中心である。
abstractThe Enviratron is an array of custom-designed plant growth chambers that simulate different environmental conditions coupled with precise sensor-based phenotypic measurements carried out by a robotic rover.
Reproduction assets foundThe paper describes the Enviratron phenotyping facility and explicitly states that a Git repository containing the code developed to operate and support the Enviratron is publicly available on GitLab. This is an authors' public code asset directly tied to this paper's phenotyping system. No phenotype datasets or image/Code · publicA git repository containing code developed to operate and support the Enviratron is available online at https://gitlab.com/dill_picl/enviratron .Open asset ↗dill_picl/enviratronlines:165-207Plant phenotyping relevance match · UnverifiedOpenAlex · checked 13 Sept 2026
This study was conducted to estimate the dimensions of horticultural products and the mean plant height of plug seedlings using three-dimensional (3D) images.Two types of camera, a ToF camera and a stereo-vision camera, were used to acquire 3D images for horticultural products and plug seedlings.The errors calculated from the ToF images for dimensions of horticultural products and mean height of plug seedlings were lower than those predicted from stereo-vision images.A new indicator was defined for determining the mean plant height of plug seedlings.Except for watermelon with tap, the errors of circumference and height of horticultural products were 0.0-3.0%and 0.0-4.7%,respectively.Also, the error of mean plant height for plug seedlings was 0.0-5.5%.The results revealed that 3D images can be utilized to estimate accurately the dimensions of horticultural products and the plant height of plug seedlings.Moreover, our method is potentially applicable for segmenting objects and for removing outliers from the point cloud data based on the 3D images of horticultural crops.
Why it matches plant phenotyping methods3D画像を用いて園芸作物の寸法と苗の平均草丈を推定する手法を開発・誤差検証しており、植物形質の取得・抽出が研究の中心です。
abstractThis study was conducted to estimate the dimensions of horticultural products and the mean plant height of plug seedlings using three-dimensional (3D) images.
In-field citrus detection and localisation are highly challenging tasks due to varying illumination conditions, partial occlusion of citrus, and the colour variation of citrus at different stages of maturity. A reliable algorithm based on red-green-blue-depth (RGB-D) images was developed to detect and locate citrus in real, outdoor orchard environments for robotic harvesting. A depth filter and a Bayes-classifier-based image segmentation method were first developed to exclude as many backgrounds as possible. A density clustering method was then used to group adjacent points in the filtered RGB-D images into clusters, where each cluster represents a possible citrus. A colour, gradient, and geometry feature-based support vector machine classifier was trained to remove false positives. To test the method, a dataset with 506 RGB-D images was acquired in a citrus orchard on sunny and cloudy days. Results showed that the proposed algorithm was robust with an F1 score of 0.9197; the positioning errors in the x, y and z directions were 7.0 ± 2.5 mm, −4.0 ± 3.0 mm and 13.0 ± 3.0 mm, respectively, and the sizing error was −1.0 ± 4.0 mm. These excellent performance values demonstrate that the proposed method could be used to guide a citrus-harvesting robot.
Why it matches plant phenotyping methodsRGB-D画像から柑橘果実を検出・位置決めし、さらに果実サイズを推定する手法を開発・検証しており、収穫対象の単なる位置検出を超えた器官形質測定が中心です。
abstractA reliable algorithm based on red-green-blue-depth (RGB-D) images was developed to detect and locate citrus in real, outdoor orchard environments for robotic harvesting.
Plant morphological data are an important basis for precision agriculture and plant phenomics. The three-dimensional (3D) geometric shape of plants is complex, and the 3D morphology of a plant changes relatively significantly during the full growth cycle. In order to make high-throughput measurements of the 3D morphological data of greenhouse plants, it is necessary to frequently adjust the relative position between the sensor and the plant. Therefore, it is necessary to frequently adjust the Kinect sensor position and consequently recalibrate the Kinect sensor during the full growth cycle of the plant, which significantly increases the tedium of the multiview 3D point cloud reconstruction process. A high-throughput 3D rapid greenhouse plant point cloud reconstruction method based on autonomous Kinect v2 sensor position calibration is proposed for 3D phenotyping greenhouse plants. Two red–green–blue–depth (RGB-D) images of the turntable surface are acquired by the Kinect v2 sensor. The central point and normal vector of the axis of rotation of the turntable are calculated automatically. The coordinate systems of RGB-D images captured at various view angles are unified based on the central point and normal vector of the axis of the turntable to achieve coarse registration. Then, the iterative closest point algorithm is used to perform multiview point cloud precise registration, thereby achieving rapid 3D point cloud reconstruction of the greenhouse plant. The greenhouse tomato plants were selected as measurement objects in this study. Research results show that the proposed 3D point cloud reconstruction method was highly accurate and stable in performance, and can be used to reconstruct 3D point clouds for high-throughput plant phenotyping analysis and to extract the morphological parameters of plants.
Why it matches plant phenotyping methodsKinect RGB-Dによる植物の3D点群再構成と形態パラメータ抽出を中心に、センサー自己校正と高スループット化を開発・評価しているため。
abstractA high-throughput 3D rapid greenhouse plant point cloud reconstruction method based on autonomous Kinect v2 sensor position calibration is proposed for 3D phenotyping greenhouse plants.
Non-destructive plant growth parameters measurement is an important concern in automatic-seedling transplanting. Recently, several image-based monitoring approaches have been proposed and potentially developed for several agricultural applications. The presented study proposed and developed a RealSense-based machine vision system for the close-shot seedling-lump integrated monitoring. The strategy was based on the close-shot depth information. Further, the point cloud clustering and suitable algorithms were applied to obtain the segmentation of 3D seedling models. In addition, the data processing pipeline was developed to assess the different morphological parameter of 4 different seedling varieties. The experiments were carried out with 4 different seedling varieties (pepper, tomato, cucumber, and lettuce) and trained under different light conditions (light and dark). Moreover, analysis results showed that there was not significantly different (p < 0.05) found towards light and dark environments due to close-shot near-infrared detection. However, the results revealed that the stem diameter relationship between RealSense and the manual method was found for R2 = 0.68 cucumber, R2 = 0.54 tomato, R2 = 0.35 pepper, and R2 = 0.58 lettuce seedlings. Whereas, the seedling height relationship between RealSense and the manual method was found higher than R2 = 0.99, 0.99, 0.99, and 0.99 for pepper, tomato, cucumber, and lettuce, respectively. Based on the experiment results, it was concluded that the RGB-D integrated monitoring system with the purposed method could be practiced for nursery seedlings most promisingly without high labour requirements in terms of ease of use. The system revealed a good sturdiness and relevance for plant growth monitoring. Additionally, it has the perspective for future practical value to real-time vision servo operations for transplanting robots.
Why it matches plant phenotyping methodsRealSense深度カメラ、点群クラスタリング、処理パイプラインを開発し、苗の形態形質を非破壊測定して手動法と検証しているため、植物フェノタイピング手法が中心である。
abstractThe presented study proposed and developed a RealSense-based machine vision system for the close-shot seedling-lump integrated monitoring.
Vineyard yield estimation provides the winegrower with insightful information regarding the expected yield, facilitating managerial decisions to achieve maximum quantity and quality and assisting the winery with logistics. The use of proximal remote sensing technology and techniques for yield estimation has produced limited success within viticulture. In this study, 2-D RGB and 3-D RGB-D (Kinect sensor) imagery were investigated for yield estimation in a vertical shoot positioned (VSP) vineyard. Three experiments were implemented, including two measurement levels and two canopy treatments. The RGB imagery (bunch- and plant-level) underwent image segmentation before the fruit area was estimated using a calibrated pixel area. RGB-D imagery captured at bunch-level (mesh) and plant-level (point cloud) was reconstructed for fruit volume estimation. The RGB and RGB-D measurements utilised cross-validation to determine fruit mass, which was subsequently used for yield estimation. Experiment one's (laboratory conditions) bunch-level results achieved a high yield estimation agreement with RGB-D imagery (r 2 = 0.950), which outperformed RGB imagery (r 2 = 0.889). Both RGB and RGB-D performed similarly in experiment two (bunch-level), while RGB outperformed RGB-D in experiment three (plant-level). The RGB-D sensor (Kinect) is suited to ideal laboratory conditions, while the robust RGB methodology is suitable for both laboratory and in-situ yield estimation.
Why it matches plant phenotyping methodsブドウの収量という植物形質を、RGB/RGB-D画像、画像分割、3D再構成、校正、交差検証で推定する手法を比較・評価しており、フェノタイピング手法が中心です。
abstractThe use of proximal remote sensing technology and techniques for yield estimation has produced limited success within viticulture.