Research On Gongger Grass LAI Based On The Image Of HSI And Hyperion

Leaf area index (LAI) is to determine the ecosystem of the land surface of matter and energy exchange of the size of one of the structural parameters, estimating crop yields, plant growth monitoring, pest and disease detection plays an irreplaceable role.This study is based on the measured data of Gongger grass to establish a new inversion model applicable respectively in the HSI images and Hyperion image grassland LAI inversion by finding sensitive band to build a new type of inversion model to improve the traditional vegetation index inversion ineffective phenomenon, and compare the pros and cons of the different model inversion results, the optimal model of higher precision and more practical, and through the establishment of the neural network to improve the retrieval accuracy, add a new way grassland to LAI inversion. The results show that the optimal model for HSI image of the quadratic polynomial model, LAI=-115.26*R7712+81.392*R771-12.889, and the coefficient of determination is0.8911; optimal Hyperion image model as a Gaussian function model, the optimal model hyper-spectral images for Gongger grass LAI=1.287*e-(R395-0.2313)/0.002877)2+141*e-((R895-0.1653)/0.07217)2+0.8635*e-(R895-0.2634)/0.01375)2,the coefficient of determination is0.6506, they are the optimal model of LAI inversion in this area.With using the C++language I design and implementation related algorithms to construct the leaf area index inversion procedure, the formation of LAI inversion operational processes, improve production efficiency and to lay the foundation for a large area of application of the leaf area index.