AUTHOR=Song Peng , Li Zhengda , Yang Meng , Shao Yang , Pu Zhen , Yang Wanneng , Zhai Ruifang 

TITLE=Dynamic detection of three-dimensional crop phenotypes based on a consumer-grade RGB-D camera

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2023.1097725

DOI=10.3389/fpls.2023.1097725

ISSN=1664-462X

ABSTRACT=Nondestructive detection of crop phenotypic traits in the field is very important for crop breeding. Ground-based mobile platform equipped with sensors can obtain crop phenotypic traits efficiently and accurately. In this study, we proposed a dynamic field 3D phenotyping method suitable for various crops by using a consumer-grade RGB-D camera installed on a ground-based movable platform, which can collect RGB image as well as depth image of crop canopy sequence dynamically. Scale-invariant feature transform (SIFT) operator was used to detect adjacent date frame acquired by the RGB-D camera to calculate the point cloud alignment coarse matching matrix and the displacement distance of adjacent images. And the data frames used for point cloud matching were selected according to the calculated displacement distance. Then Colored ICP(Iterative Closest Point) algorithm was used to determine the fine matching matrix and generate point clouds of crop row. Clustering method was applied to segment the point cloud of each plant from crop row point cloud and 3D phenotypic traits including plant height, leaf area and projected area of individual plant were measured. We compared the effects of LIDAR and image-based 3D reconstruction methods, and experiments were carried out on and corn, tobacco, potato and bletilla striata in seedling stage. The results showed that the measurement of plant height (R²=0.9~0.96, RSME = 0.015~0.022m), leaf area (R²= 0.8~0.85, RSME = 0.0011~0.0041m^2) and projected area (R² = 0.96~0.99) have strong correlation with manual measurement results. Meanwhile, 3D reconstruction results with different moving speeds and with different period of a day were also verified. Results showed that the method can be applied to dynamic detection with a moving speed up to 0.6m/s, and can achieve acceptable detection results in daytime as well as at night. Thus, the proposed method can improve the efficiency of individual crop 3D point cloud data extraction within acceptable accuracy, which is a feasible solution for crop seedling 3D phenotyping outdoor.