The development of computer vision-based rice phenotyping techniques is crucial for precision field manage-ment and accelerated breeding,which facilitate continuously advancing rice production.Among phenotyping tasks,...The development of computer vision-based rice phenotyping techniques is crucial for precision field manage-ment and accelerated breeding,which facilitate continuously advancing rice production.Among phenotyping tasks,distinguishing image components is a key prerequisite for characterizing plant growth and development at the organ scale,enabling deeper insights into ecophysiological processes.However,owing to the fine structure of rice organs and complex illumination within the canopy,this task remains highly challenging,underscoring the need for a high-quality training dataset.Such datasets are scarce,both because of a lack of large,representative collections of rice field images and because of the time-intensive nature of the annotation.To address this gap,we created the first comprehensive multiclass rice semantic segmentation dataset,RiceSEG.We gathered nearly 50,000 high-resolution,ground-based images from five major rice-growing countries(China,Japan,India,the Philippines,and Tanzania),encompassing more than 6000 genotypes across all growth stages.From these original images,3078 representative samples were selected and annotated with six classes(background,green vegetation,senescent vegetation,panicle,weeds,and duckweed)to form the RiceSEG dataset.Notably,the subdataset from China spans all major genotypes and rice-growing environments from northeastern to southern regions.Both state-of-the-art convolutional neural networks and transformer-based semantic segmentation models were used as baselines.While these models perform reasonably well in segmenting background and green vegetation,they face difficulties during the reproductive stage,when canopy structures are more complex and when multiple classes are involved.These findings highlight the importance of our dataset for developing specialized segmentation models for rice and other crops.The RiceSEG dataset is publicly available at .展开更多
The nondestructive and rapid acquisition of rice field phenotyping information is very important for the precision management of the rice growth process.In this research,the phenotyping information LAI(leaf area index...The nondestructive and rapid acquisition of rice field phenotyping information is very important for the precision management of the rice growth process.In this research,the phenotyping information LAI(leaf area index),leaf chlorophyll content(C_(ab)),canopy water content(C_(w)),and dry matter content(C_(dm))of rice was inversed based on the hyperspectral remote sensing technology of an unmanned aerial vehicle(UAV).The improved Sobol global sensitivity analysis(GSA)method was used to analyze the input parameters of the PROSAIL model in the spectral band range of 400-1100 nm,which was obtained by hyperspectral remote sensing by the UAV.The results show that C_(ab) mainly affects the spectrum on 400-780 nm band,C_(dm) on 760-1000 nm band,C_(w) on 900-1100 nm band,and LAI on the entire band.The hyperspectral data of the 400-1100 nm band of the rice canopy were acquired by using the M600 UAV remote sensing platform,and the radiance calibration was converted to the canopy emission rate.In combination with the PROSAIL model,the particle swarm optimization algorithm was used to retrieve rice phenotyping information by constructing the cost function.The results showed the following:(1)an accuracy of R^(2)=0.833 and RMSE=0.0969,where RMSE denotes root-mean-square error,was obtained for C_(ab) retrieval;R^(2)=0.816 and RMSE=0.1012 for LAI inversion;R^(2)=0.793 and RMSE=0.1084 for C_(dm);and R^(2)=0.665 and RMSE=0.1325 for C_(w).The C_(w) inversion accuracy was not particularly high.(2)The same band will be affected by multiple parameters at the same time.(3)This study adopted the rice phenotyping information inversion method to expand the rice hyperspectral information acquisition field of a UAV based on the phenotypic information retrieval accuracy using a high level of field spectral radiometric accuracy.The inversion method featured a good mechanism,high universality,and easy implementation,which can provide a reference for nondestructive and rapid inversion of rice biochemical parameters using UAV hyperspectral remote sensing.展开更多
Advanced plant phenotyping technologies are vital for trait improvement and accelerating intelligent breeding.Due to the species diversity of plants,existing methods heavily rely on large-scale high-precision manually...Advanced plant phenotyping technologies are vital for trait improvement and accelerating intelligent breeding.Due to the species diversity of plants,existing methods heavily rely on large-scale high-precision manually annotated data.For self-occluded objects at the grain level,unsupervised methods often prove ineffective.This study proposes IPENS,an interactive unsupervised multi-target point cloud extraction method.It utilizes radi-ance field information to lift 2D masks,segmented by SAM2(Segment Anything Model 2),into 3D space for target point cloud extraction.A multi-target collaborative optimization strategy addresses the challenge of segmenting multiple targets from a single interaction.On a rice dataset,IPENS achieves a grain-level segmen-tation mean Intersection over Union(mIoU)of 63.72%.For phenotypic trait estimation,it achieves a grain voxel volume coefficient of determination R^(2)=0.7697(Root Mean Square Error,RMSE=0.0025),leaf surface area R^(2)=0.84(RMSE=18.93),and leaf length and width prediction accuracies of R^(2)=0.97 and R^(2)=0.87(RMSE=1.49 and 0.21).On a wheat dataset,IPENS further improves segmentation performance to a mIoU of 89.68%,with exceptional phenotypic estimation results:panicle voxel volume R^(2)=0.9956(RMSE=0.0055),leaf surface area R^(2)=1.00(RMSE=0.67),and leaf length and width predictions reaching R^(2)=0.99 and R^(2)=0.92(RMSE=0.23 and 0.15).Without requiring annotated data,IPENS rapidly extracts grain-level point clouds for multiple targets within 3 min using single-round image interactions.These features make IPENS a high-quality,non-invasive phenotypic extraction solution for rice and wheat,offering significant potential to enhance intelligent breeding.展开更多
基金This work was supported by the National Key R&D Program of China(No.2022YFE0116200 and No.2022YFD2300700)the Young Scientists Fund of the National Natural Science Foundation of China(No.42201437 and No.32201893)+4 种基金the PhD Scientific Research and Innovation Foundation of The Education Department of Hainan Province Joint Project of Sanya Yazhou Bay Science and Technology City(No.HSPHDSRF-2024-09-001)the Hainan Provincial Natural Science Foundation of China(No.325QN370)the“JBGS”Project of Seed Industry Revitalization in Jiangsu Province(JBGS[2021]007)the Japan Society for the Promotion of Science(No.22KK0083 and No.JP25H01110)the Sarabetsu Village“Endowed Chair for Field Phenomics”project in Hokkaido,Japan.
文摘The development of computer vision-based rice phenotyping techniques is crucial for precision field manage-ment and accelerated breeding,which facilitate continuously advancing rice production.Among phenotyping tasks,distinguishing image components is a key prerequisite for characterizing plant growth and development at the organ scale,enabling deeper insights into ecophysiological processes.However,owing to the fine structure of rice organs and complex illumination within the canopy,this task remains highly challenging,underscoring the need for a high-quality training dataset.Such datasets are scarce,both because of a lack of large,representative collections of rice field images and because of the time-intensive nature of the annotation.To address this gap,we created the first comprehensive multiclass rice semantic segmentation dataset,RiceSEG.We gathered nearly 50,000 high-resolution,ground-based images from five major rice-growing countries(China,Japan,India,the Philippines,and Tanzania),encompassing more than 6000 genotypes across all growth stages.From these original images,3078 representative samples were selected and annotated with six classes(background,green vegetation,senescent vegetation,panicle,weeds,and duckweed)to form the RiceSEG dataset.Notably,the subdataset from China spans all major genotypes and rice-growing environments from northeastern to southern regions.Both state-of-the-art convolutional neural networks and transformer-based semantic segmentation models were used as baselines.While these models perform reasonably well in segmenting background and green vegetation,they face difficulties during the reproductive stage,when canopy structures are more complex and when multiple classes are involved.These findings highlight the importance of our dataset for developing specialized segmentation models for rice and other crops.The RiceSEG dataset is publicly available at .
基金support of the National Key Research and Development Plan of China(Grant No.2016YFD020060307)Key Project of Education Department of Liaoning province(LSNZD201605).
文摘The nondestructive and rapid acquisition of rice field phenotyping information is very important for the precision management of the rice growth process.In this research,the phenotyping information LAI(leaf area index),leaf chlorophyll content(C_(ab)),canopy water content(C_(w)),and dry matter content(C_(dm))of rice was inversed based on the hyperspectral remote sensing technology of an unmanned aerial vehicle(UAV).The improved Sobol global sensitivity analysis(GSA)method was used to analyze the input parameters of the PROSAIL model in the spectral band range of 400-1100 nm,which was obtained by hyperspectral remote sensing by the UAV.The results show that C_(ab) mainly affects the spectrum on 400-780 nm band,C_(dm) on 760-1000 nm band,C_(w) on 900-1100 nm band,and LAI on the entire band.The hyperspectral data of the 400-1100 nm band of the rice canopy were acquired by using the M600 UAV remote sensing platform,and the radiance calibration was converted to the canopy emission rate.In combination with the PROSAIL model,the particle swarm optimization algorithm was used to retrieve rice phenotyping information by constructing the cost function.The results showed the following:(1)an accuracy of R^(2)=0.833 and RMSE=0.0969,where RMSE denotes root-mean-square error,was obtained for C_(ab) retrieval;R^(2)=0.816 and RMSE=0.1012 for LAI inversion;R^(2)=0.793 and RMSE=0.1084 for C_(dm);and R^(2)=0.665 and RMSE=0.1325 for C_(w).The C_(w) inversion accuracy was not particularly high.(2)The same band will be affected by multiple parameters at the same time.(3)This study adopted the rice phenotyping information inversion method to expand the rice hyperspectral information acquisition field of a UAV based on the phenotypic information retrieval accuracy using a high level of field spectral radiometric accuracy.The inversion method featured a good mechanism,high universality,and easy implementation,which can provide a reference for nondestructive and rapid inversion of rice biochemical parameters using UAV hyperspectral remote sensing.
基金supported by the National Key Research and Development Program of China(Grant Number 2023YFD1901003)the Strategic Priority Research Program of the Chinese Academy of Sciences(Grant XDA28120402).
文摘Advanced plant phenotyping technologies are vital for trait improvement and accelerating intelligent breeding.Due to the species diversity of plants,existing methods heavily rely on large-scale high-precision manually annotated data.For self-occluded objects at the grain level,unsupervised methods often prove ineffective.This study proposes IPENS,an interactive unsupervised multi-target point cloud extraction method.It utilizes radi-ance field information to lift 2D masks,segmented by SAM2(Segment Anything Model 2),into 3D space for target point cloud extraction.A multi-target collaborative optimization strategy addresses the challenge of segmenting multiple targets from a single interaction.On a rice dataset,IPENS achieves a grain-level segmen-tation mean Intersection over Union(mIoU)of 63.72%.For phenotypic trait estimation,it achieves a grain voxel volume coefficient of determination R^(2)=0.7697(Root Mean Square Error,RMSE=0.0025),leaf surface area R^(2)=0.84(RMSE=18.93),and leaf length and width prediction accuracies of R^(2)=0.97 and R^(2)=0.87(RMSE=1.49 and 0.21).On a wheat dataset,IPENS further improves segmentation performance to a mIoU of 89.68%,with exceptional phenotypic estimation results:panicle voxel volume R^(2)=0.9956(RMSE=0.0055),leaf surface area R^(2)=1.00(RMSE=0.67),and leaf length and width predictions reaching R^(2)=0.99 and R^(2)=0.92(RMSE=0.23 and 0.15).Without requiring annotated data,IPENS rapidly extracts grain-level point clouds for multiple targets within 3 min using single-round image interactions.These features make IPENS a high-quality,non-invasive phenotypic extraction solution for rice and wheat,offering significant potential to enhance intelligent breeding.
