Plant Phenomics最新文献

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 R² of 0.99 compared with manual measurements. Similar experiments with rapeseed yielded an R² 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.

Accurate crop disease classification is essential for disease management to support food security. Deep learning has shown its high classification accuracy in image-based disease identification. However, the deep learning approach usually needs large amounts of data for training to achieve satisfactory performance, which hindering its application and scalability for different crops. Large language models (LLMs) have shown strong generation capability and zero-shot performance. While how to utilize the LLM technique for crop disease classification remains unclear. In this study, we developed a training-free framework named ChatLeafDisease (ChatLD) based on GPT-4o model with chain-of-thought (CoT) prompting for crop disease classification. The framework includes a disease description database to provide knowledge of crop diseases and a disease classification agent guided by CoT prompts to understand the patterns of leaves infected diseases and classify the disease. The original GPT-4o model, Gemini model, and Contrastive Language-Image Pre-training (CLIP) model were chosen as baselines. Results showed that the ChatLD framework achieved higher and more stable classification accuracy (88.9 ​%) for six tomato diseases than the GPT-4o (45.9 ​%), Gemini (56.1%), and CLIP (64.3 ​%) models. We found that the scoring rules enabled the ChatLD framework to capture the typical differences across diseases. Ablation results showed that the CoT prompts integrated the scoring rules and important notes to enable the ChatLD to achieve high classification accuracy. Comparison between different description texts showed that condensed disease description improved the classification performance. The results showed that the ChatLD framework achieved high accuracy for the disease classes of new crops, highlighting its scalability across various crop diseases. The proposed framework provided a new LLM-based alternative for crop disease classification by only using the textual descriptions of disease without training process.

Computer vision is increasingly used in farmers' fields and agricultural experiments to quantify important traits. Imaging setups with a sub-millimeter ground sampling distance enable the detection and tracking of plant features, including size, shape, and colour. Although today's AI-driven foundation models segment almost any object in an image, they still fail for complex plant canopies. To improve model performance, the global wheat dataset consortium assembled a diverse set of images from experiments around the globe. After the head detection dataset (GWHD), the new dataset targets a full semantic segmentation (GWFSS) of organs (leaves, stems and spikes) covering all developmental stages. Images were collected by 11 institutions using a wide range of imaging setups. Two datasets are provided: i) a set of 1096 diverse images in which all organs were labelled at the pixel level, and (ii) a dataset of 52,078 images without annotations available for additional training. The labelled set was used to train segmentation models based on DeepLabV3Plus and Segformer. Our Segformer model performed slightly better than DeepLabV3Plus with a mIOU for leaves and spikes of ca. 90 ​%. However, the precision for stems with 54 ​% was rather lower. The major advantages over published models are: i) the exclusion of weeds from the wheat canopy, ii) the detection of all wheat features including necrotic and senescent tissues and its separation from crop residues. This facilitates further development in classifying healthy vs. unhealthy tissue to address the increasing need for accurate quantification of senescence and diseases in wheat canopies.

Anthesis prediction is crucial for breeding wheat. While current tools provide estimates of average anthesis at the field scale, they fail to address the needs of breeders who require accurate predictions for individual plants. Hybrid breeders have to finalize their plans for pollination at least 10 days before such flowering is due and biotechnology field trials in the United States and Australia must report to regulators 7-14 days before the first plant flowers. Currently, predicting anthesis of individual wheat plants is a labour-intensive, inefficient, and costly process. Individual wheat of the same cultivar within the same field may exhibit substantial variations in anthesis timing, due to significant variations in their immediate surroundings. In this study, we developed an efficient and cost-effective machine vision approach to predict anthesis of individual wheat plants. By integrating RGB imagery with in-situ meteorological data, our multimodal framework simplifies the anthesis prediction problem into binary or three-class classification tasks, aligning with breeders' requirements in individual wheat flowering prediction on the crucial days before anthesis. Furthermore, we incorporated a few-shot learning method to improve the model's adaptability across different growth environments and to address the challenge of limited training data. The model achieved an F1 score above 0.8 in all planting settings.

As an important tropical cash crop, rubber trees play a key role in the rubber industry and ecosystem. However, a significant challenge in precision agriculture and refined management of rubber plantation lies in the limitations of traditional point cloud segmentation methods, which struggle to accurately extract structural parameters and capture the spatial layout of individual rubber trees. Therefore, we propose an optimized dual-channel clustering method for the UAV LiDAR-based Rubber Tree Point Cloud Segmentation Network (RsegNet) for improved assessment of rubber tree architecture and traits. Firstly, we designed a cosine feature extraction network, termed CosineU-Net, to address the branch-and-leaf overlap problem by calculating the cosine similarity of the spatial and positional features of each point, leveraging deep learning approaches to improve feature representation. Secondly, we constructed a dual-channel clustering module reducing prediction error in rubber tree point cloud data, integrating multi-class association and background classification to tackle background interference. The cluster identification and separation accuracy in high-dimensional data processing is enhanced through a dynamic clustering optimization algorithm. In our self-built dataset and across five regions of the FOR-instance forest dataset, RsegNet achieved the best performance compared to five state-of-the-art networks, reaching an F-score of 86.1%. This method calculated structural attributes including height, crown diameter, and volume for rubber trees in three areas under different environments in Danzhou City, Hainan Province, providing robust support for precise monitoring, plantation management, and health assessment.

Accurate predictions and representations of plant growth patterns in simulated and controlled environments are important for addressing various challenges in plant phenomics research. This review explores various works on state-of-the-art predictive pattern recognition techniques, focusing on the spatiotemporal modeling of plant traits and the integration of dynamic environmental interactions. We provide a comprehensive examination of deterministic, probabilistic, and generative modeling approaches, emphasizing their applications in high-throughput phenotyping and simulation-based plant growth forecasting. Key topics include regressions and neural network-based representation models for the task of forecasting, limitations of existing experiment-based deterministic approaches, and the need for dynamic frameworks that incorporate uncertainty and evolving environmental feedback. This review surveys advances in 2D and 3D structured data representations through functional-structural plant models and conditional generative models. We offer a perspective on opportunities for future works, emphasizing the integration of domain-specific knowledge to data-driven methods, improvements to available datasets, and the implementation of these techniques toward real-world applications.

A novel approach, the Algorithmic Root Trait (ART) extraction method, identifies and quantifies computationally-derived plant root traits, revealing latent patterns related to dense root clusters in digital images. Using an ensemble of multiple unsupervised machine learning algorithms and a custom algorithm, 27 ARTs were extracted reflecting dense root cluster size and spatial location. These ARTs were then used independently and in combination with Traditional Root Traits (TRTs) to classify wheat genotypes differing in drought tolerance. ART-based models outperformed TRT-only models in drought classification (e.g., 96.3 ​% vs. 85.6 ​% accuracy). Combining ARTs and TRTs further improved accuracy to 97.4 ​%. Notably, 4 selected ARTs matched the performance of all 23 TRTs, offering 5.8 ​× ​higher information density (0.213 vs. 0.037 accuracy/feature). This superiority reflects the ability of ARTs to capture richer, more complex architectural information, evidenced by higher internal variability (35.59 ​± ​11.41 vs. 28.91 ​± ​14.28 for TRTs) and distinct data structures in multivariate analyses; PERMANOVA confirmed that ARTs and TRTs provide complementary insights. Validated through experiments in controlled environments and field conditions with wheat drought-tolerant and susceptible genotypes, ART offers a scalable, customisable toolset for high-throughput phenotyping of plant roots. By bridging conventional, visually derived traits with autonomous computational analyses, this method broadens root phenotyping pipelines and underscores the value of harnessing sensor data that transcends human perception. ART thus emerges as a promising framework for revealing hidden features in plant imaging, with broader applications across plant science to deepen our understanding of crop adaptation and resilience.

Deep learning-based crop and weed detection is essential for modern precision weed control. But its effectiveness is limited when facing newly presented weed species due to the impracticality of collecting large, balanced training datasets in field conditions. To address these challenges, this study presents a few-shot learning framework that achieves rapid and effective adaptation to new weed species by leveraging domain-specific characteristics of plant detection. We proposed few-shot enhanced attention (FSEA) network, built upon Faster R-CNN, which implements three prior knowledge in weed detection through: (1) designing a channel attention-based feature fusion module with an excess-green feature extractor to leverage color characteristics of plants and background, (2) designing a feature enhancement module to accommodate diverse plant morphologies, and (3) applying an optimized loss function designed specifically for plant occlusion scenarios. Using commonly observed crop and weed species (common beet, sugarcane, barnyard grass, field pennycress and Chinese money plant) as base classes, FSEA achieved an all-class mAP of 0.416 and a novel-class mAP of 0.346 when adapting to less frequent weed species (common purslane, Asian copperleaf, goosefoot, clover, and goosegrass), after training for 40 epochs using only 30 samples per species. This performance significantly outperforms state-of-the-art few-shot detectors (TFA, FSCE, Meta R-CNN, Meta-DETR, DCFS, DiGEO) and traditional detector YOLOv7, indicating the effectiveness of incorporating domain-specific prior knowledge into few-shot weed detection. This study provides a fundamental methodology for rapid adaptation of weed detection systems to new environments and species, making automated weed management more practical and accessible for various agricultural applications. The source code and dataset are publicly available (https://github.com/skyofyao/FSEA) to facilitate further research in this domain.

Metabolic processes in plant organs involving transport of water, metabolic gasses, and nutrients depend on the three-dimensional (3D) microscopic tissue morphology. However, imaging and quantifying this microstructure, including the spatial layout of parenchyma cells, pores, vascular bundles and special features such as stone cell clusters (brachysclereids), is challenging. To address this, a 3D deep learning-based panoptic segmentation model, combining semantic and instance segmentation, was developed to accelerate and improve microstructure characterization of apple and pear fruit tissue in X-ray micro-computed tomography (CT) images. In addition, various training datasets and data augmentation techniques, including synthetic data, were explored to enhance segmentation quality. The 3D panoptic segmentation achieved an Aggregated Jaccard Index of 0.89 and 0.77 for apple and pear tissue, respectively, outperforming both the previously designed 2D instance segmentation model and a marker-based watershed segmentation benchmark. The model successfully labeled vascular bundles with a Dice Similarity Coefficient (DSC) of 0.51 in apple tissue and 0.79 in pear tissue, although thin vasculature in apple remained more challenging to segment. The 3D panoptic segmentation model achieved a DSC of 0.81 and effectively segmented stone cell clusters in pear tissue. Despite evaluating different methods to enhance segmentation quality, none improved test performance beyond that of the model trained on the standard dataset. The proposed 3D panoptic segmentation model offers the most complete automated protocol to date for plant tissue labelling and morphometric quantification from native X-ray micro-CT images, without extensive sample preparation such as contrast labelling. The developed method, if not replaces, drastically accelerates conventional human-in-the-loop analysis of such images.

3D phenotyping of the external and internal structures is important to breed new fruit species. As manual phenotyping is error-prone and time-consuming, developing high-throughput solutions with enhanced precision and low costs is necessary. This study presents CitrusGAN, a generative adversarial network-based method to reconstruct 3D citrus CT models from sparse-view X-ray images. The input X-rays are arranged in orthogonal pairs to provide additional information, and customized loss functions enable more effective learning of the mapping from 2D X-ray features to 3D CT volumes. Experimental results show that 6 views can generate high-quality citrus CT volumes, with a structural similarity index of 92.1 ​% and a peak signal-to-noise ratio of 26.374 ​dB compared with the real CT models. Moreover, the morphology of the generated model can be conveniently measured in the 3D space, facilitating the extraction of phenotypic traits including fruit length, width, height, volume, surface area, peel thickness, number of segments, and edible rate with high precision. As X-rays can be obtained using low-cost X-ray machines with high efficiency, the proposed method can be potentially developed into high-throughput equipment for fruit production lines or portable devices to realize in-field phenotyping.