Estimation Of Vegetation Coverage In Hyperspectral Images Based On Improved Pixel Dichotomy

Vegetation coverage,as a type of land coverage,is an important component of the ecosystem,an important parameter representing the amount of vegetation on the land surface,and an important indicator reflecting the change of the ecosystem.Inversion of surface biophysical parameters by remote sensing has been more and more widely used in all works of life,so it has also become the main means of Fractional vegetation cover measurement and monitoring.However,at present,it mainly uses multi-spectral remote sensing data to extract Fractional vegetation cover information in a wide range.With the rapid development of hyperspectral resolution remote sensing technology,remote sensing can directly carry out the quantitative analysis of weak spectral differences on ground objects,which also provides the possibility for the inversion and extraction of Fractional vegetation cover with higher accuracy.This paper uses Pixel dichotomy to Hyperion hyperspectral remote sensing image,the key technology of vegetation coverage inversion research development based on the combination of genetic algorithm optimization evaluation model of NDVI band selection study,and based on the optimal density segmentation threshold segmentation research,to find a suitable method for Fractional vegetation cover of hyperspectral data feature extraction,and provide technical support for experimental area ecological environment evaluation and precise regulation.The main research contents and achievements are as follows:(1)According to the traditional method to determine the hyperspectral calculation of NDVI,on the basis of analyzing the commonly used band selection criteria,the central band is used as the R and NIR band method.A combination evaluation model based on genetic algorithm to optimize the weight coefficient is proposed for NDVI band selection,and the band combinations that meet the requirements of various evaluation indexes are selected.(2)The threshold segmentation method is introduced to study the value of NDVI_min and NDVI_max in the pixel binary model.The optimal density segmentation method is used to segment different segments of the image,and the best segment number is selected,so as to determine the NDVI_min and NDVI_max thresholds of various ground objects.(3)An improved algorithm of pixel dichotomy was proposed to improve the inversion accuracy of vegetation coverage by first classifying ground features and then extracting vegetation coverage.At the same time,the traditional pixel dichotomy method was used to obtain the vegetation coverage of the study area to form a comparative experiment and verify the results.Experimental results show that the improved algorithm can improve the extraction accuracy of vegetation coverage,and it is suitable for both high and low vegetation coverage areas.However,the improved algorithm has better applicability in areas with higher vegetation coverage...