Search for: All records

Understanding the complex interactions between genotype-environment dynamics is fundamental for optimizing crop improvement. However, traditional phenotyping methods limit assessments to the end of the growing season, restricting continuous crop monitoring. To address this limitation, we developed a methodology for spatiotemporal registration of time-series 3D point cloud data, enabling field phenotyping over time for accurate crop growth tracking. Leveraging multi-scan terrestrial laser scanning (TLS), we captured high-resolution 3D LiDAR data in a cotton breeding field across various stages of the growing season to generate four-dimensional (4D) crop models, seamlessly integrating spatial and temporal dimensions. Our registration procedure involved an initial pairwise terrain-based matching for rough alignment, followed by a bird’s-eye view adjustment for fine registration. Point clouds collected throughout nine sessions across the growing season were successfully registered both spatially and temporally, with average registration errors of approximately 3 cm. We used the generated 4D models to monitor canopy height (CH) and volume (CV) for eleven cotton genotypes over two months. The consistent height reference established via our spatiotemporal registration process enabled precise estimations of CH (R² = 0.95, RMSE = 7.6 cm). Additionally, we analyzed the relationship between CV and the interception of photosynthetically active radiation (IPAR_f), finding that it followed a curve with exponential saturation, consistent with theoretical models, with a standard error of regression (SER) of 11%. In addition, we compared mathematical models from the Richards family of sigmoid curves for crop growth modeling, finding that the logistic model effectively captured CH and CV evolution, aiding in identifying significant genotype differences. Our novel TLS-based digital phenotyping methodology enhances precision and efficiency in field phenotyping over time, advancing plant phenomics and empowering efficient decision-making for crop improvement efforts. 

Abstract BackgroundPlant architecture can influence crop yield and quality. Manual extraction of architectural traits is, however, time-consuming, tedious, and error prone. The trait estimation from 3D data addresses occlusion issues with the availability of depth information while deep learning approaches enable learning features without manual design. The goal of this study was to develop a data processing workflow by leveraging 3D deep learning models and a novel 3D data annotation tool to segment cotton plant parts and derive important architectural traits. ResultsThe Point Voxel Convolutional Neural Network (PVCNN) combining both point- and voxel-based representations of 3D data shows less time consumption and better segmentation performance than point-based networks. Results indicate that the best mIoU (89.12%) and accuracy (96.19%) with average inference time of 0.88 s were achieved through PVCNN, compared to Pointnet and Pointnet++. On the seven derived architectural traits from segmented parts, an R²value of more than 0.8 and mean absolute percentage error of less than 10% were attained. ConclusionThis plant part segmentation method based on 3D deep learning enables effective and efficient architectural trait measurement from point clouds, which could be useful to advance plant breeding programs and characterization of in-season developmental traits. The plant part segmentation code is available athttps://github.com/UGA-BSAIL/plant_3d_deep_learning. 

Estimation of cotton yield before harvest offers many benefits to breeding programs, researchers and producers. Remote sensing enables efficient and consistent estimation of cotton yields, as opposed to traditional field measurements and surveys. The overall goal of this study was to develop a data processing pipeline to perform fast and accurate pre-harvest yield predictions of cotton breeding fields from aerial imagery using machine learning techniques. By using only a single plot image extracted from an orthomosaic map, a Support Vector Machine (SVM) classifier with four selected features was trained to identify the cotton pixels present in each plot image. The SVM classifier achieved an accuracy of 89%, a precision of 86%, a recall of 75%, and an F1-score of 80% at recognizing cotton pixels. After performing morphological image processing operations and applying a connected components algorithm, the classified cotton pixels were clustered to predict the number of cotton bolls at the plot level. Our model fitted the ground truth counts with anR²value of 0.93, a normalized root mean squared error of 0.07, and a mean absolute percentage error of 13.7%. This study demonstrates that aerial imagery with machine learning techniques can be a reliable, efficient, and effective tool for pre-harvest cotton yield prediction.