Extraction Of Typical Vegetation Abundance Information In South China Based On Remote Sensing

Fractional vegetation cover is an important basic data for describing ecosystems and an important indicator for measuring the state of vegetation on the surface.Moreover,as an important indicator of regional ecosystem environmental changes,it has great significance in the fields of ecology,hydrology and regional change.Therefore,it is particularly important to improve the accuracy of vegetation coverage information in order to cope with important national natural resources.In this paper,the southwestern part of Ningxiang City,which is covered by typical vegetation in the south,is selected as the research area.Based on the linear spectral mixing model,the extraction accuracy of vegetation abundance information was improved by means of constructing vegetation time series curve,multiple endmembers matching spectral unmixing of dominant tree species,etc.The main research contents and conclusions are as follows:(1)Construction of vegetation time series curve.By constructing the EVI time series curves of the typical vegetation(endmember)and mixed pixels of the Iandsat8 time series image,the differences of different forest types are highlighted to compensate for the lack of multi-spectral data bands and the accuracy of the mixed pixels is not accurate.(2)The dominant tree species multiple endmembers matching mixed pixel decomposition.Selecting the EVI growth curve of multiple endmembers of dominant tree species,matching the EVI time series curve to match multiple endmembers by spectral matching to determine the action area of each endmember,and using EVI time series curve instead of spectral curve to participate in linear spectrum Unmixing,that achieves the purpose of using different endmember combinations for mixed pixel decomposition.(3)Comparative analysis.The method of this paper and the traditional spectral unmixing method are used to separate the mixed pixels.Compared with the traditional method,the unmixing precision of broad-leaved forest,the unmixing precision of coniferous forest and the precision of broad-leaf forest are improved.Obviously,the effectiveness of the method is proved.(4)Increase the number of endmembers of the dominant tree species involved in the demixing of mixed pixels,and determine the optimal number of endmembers of the dominant tree species involved in the unmixing.The results show that the EVI time series curve is superior to the spectral reflectance curve in the process of mixed pixel decomposition;The limited spectral band and the phenomenon of isomerism reduce the accuracy of conventional pixel decomposition based on a combination of end-member spectra;In this paper,the multiple endmembers of the dominant tree species were selected for unmixing,and the spectral end angle(EVI time series curve)was used to math the best endmember combination for unmixing.The accuracy of the hybrid pixel decomposition has been significantly improved,which proves the effectiveness of the method.Along with the economic development in the southern region,a large number of high-grade highways have been put into construction,resulting in the forest vegetation being stressed,resulting in higher variability of the same species,and the same species of wood itself is also different.In this paper,by adding the same number of endmembers to participate in the hybrid pixel decomposition,the influence of the internal difference of the same kind of endmember on the pixel decomposition precision is weakened.