Abstract:Leaf area index (LAI) is one of the important parameters to describe the corn canopy structure. PROSAIL model is a mechanism model for retriving LAI, which can express canopy situation more truly. But the leaf angle distribution function used in PROSAIL model assumed that the leaf angle distribution is constant during whole crop growth period, and it cant reflect the actual leaf angle distribution of the corn plant. This paper studied the extracting method of the maize LAI based on the PROSAIL model, using GF-1 images and terrestrial LiDAR data. In order to get the leaf angle distribution of maize, the point cloud of maize was separated into small leaf units through voxel method, and then the surface was matched according to the point cloud in each voxel. The accurate leaf angle distribution function was acquired from statistics data of each leaf unit angle. Combining with the ellipsoid distribution function, the accurate leaf angle distribution function was got, which is used to optimize the PROSAIL model. In this research, the maize canopy LAI was retrived in Farm No.852, Heilongjiang Province, through traditional PROSAIL model and optimized PROSAIL model respectively. The main conclusion is as follows: all of the two methods of inversion of LAI have a good correlation with measured LAI as coefficient is 0.5576 and 0.8583 respectively, which proved that this model is credibly. But the result of inversed LAI based on unimproved model is low. After optimized model with TLS data, the inversion of LAI estimation accuracy was improved from 26.53% to 96.23%.Therefore, it can greatly improve the accuracy of LAI inversion by introducing the TLS point cloud data to improve crops leaf angle distribution function in PROSAIL model.
