Remote Sensing Inversion Of Forest Leaf Area Index Based On Geometric Optical Model

Forest leaf area index(LAI)is an important indicator of the growth status of forest ecosystems,and an important factor that affects the role of forests in the process of terrestrial ecosystem changes.It is used to simulate the key parameters required for the interaction between regional or global terrestrial ecosystems and the atmosphere for ecology,hydrology,climate,and biogeochemistry.At present,there is a problem of low spatial resolution and inversion accuracy for remote sensing inversion of forest leaf area index.In this thesis,the forest area of Lushuihe is used as the research area,and the4-scale geometric optical model and Sentinel-2 remote sensing data are used to carry out the inversion study of the forest leaf area index in the research area.The main results and conclusions of the research are as follows:(1)Using the reflectance data simulated by the 4-scale geometric-optical model and Sentinel-2 remote sensing data,through calculation,analysis the multiple vegetation indexes and compare the correlation between vegetation indexes of the two data sources and the forest leaf area index proved that 4-scale geometric-optical model simulation data is suitable for inversion of forest leaf area index in the study area and is superior to Sentinel-2 remote sensing data.(2)By analyzing and comparing the correlation between the vegetation index,terrain parameters and the forest leaf area index under different modeling strategies,it is determined that the slope interval modeling is the optimal modeling strategy.(3)By comparing the fitting accuracy and verification accuracy of different slope intervals based on multiple stepwise regression,partial least squares regression and random forest model,the random forest model is determined as the optimal model for forest area index inversion in the study area.The research results in this thesis show that the forest leaf area index inversion based on the 4-scale geometric-optical model has certain application potential,which can provide a reference for the forest leaf area index inversion with higher accuracy and higher spatial resolution.