Dynamic Simulation of Leaf Area Index for the Soybean Canopy Based on 3D Reconstruction
Bingxue Wei,Xiaodan Ma,Haiou Guan,Miao Yu,Chen Yang,Haotian He,Feiyi Wang,Panpan Shen
DOI: https://doi.org/10.1016/j.ecoinf.2023.102070
IF: 5.1
2023-01-01
Ecological Informatics
Abstract:Leaf area index (LAI) of the soybean canopy was an important indicator to reflect the growth and development of the soybean plant. However, the traditional LAI measurement in the field was expensive, time-consuming, and challenging to achieve high accuracy. Thus, in this study, a calculation method of LAI for the soybean canopy based on 3D reconstruction was proposed, and the dynamic simulation model of canopy LAI was established. First, northeast soybean varieties of Kangxianchong8 and Dongnong252 were taken as the research objects, and a multi-source image synchronous acquisition platform for soybean canopy based on Kinect 2.0 was constructed to obtain the canopy image data from the V3 to R7 growth periods. Second, the 3D structure of the soybean canopy was reconstructed by conditional filtering and statistical filtering. Third, a soybean LAI estimation method was established by canopy analysis. The determination coefficient R2 between the calculated value and the standard value of LAI was greater than 0.99. Finally, a dynamic simulation model of soybean LAI was established based on the Richards model and the genetic parameters of varieties. The results showed that the accuracy of the dynamic simulation model reached above 0.99, which realized the critical technology of rapid detection and dynamic simulation of LAI for the soybean canopy, and provided quantitative dynamic prediction and technical support for scientific regulation of ecology and morphology for the soybean canopy.