Extraction Of Vegetation Type In The Typical Area Of Qinghai Lake Basin Based On Multi-component Mixed Spectral Library

The Qinghai Lake Basin is a barrier to maintain ecological security in the northeast of the Tibetan Plateau,and the ecological location is very important.Carrying out vegetation type identification and extraction has important theoretical value and practical significance for the protection and management of vegetation in the Qinghai Lake Basin and monitoring of the grassland ecological environment.In this paper,a typical vegetation coverage area in the Buha River Basin of Qinghai Lake is taken as an example.Based on the vegetation and soil spectral data collected on the ground,a mixed spectral library with three abundances of vegetation,soil and shadow is used to construct a mixed spectral library in the study area,And through the five algorithms of Spectral Angle Mapping(SAM),Spectral Information Divergence(SID),SID_SAM,Spectral Correlation Mapper(SCM)and Spectral Correlation Angle(SCA),the hybrid spectral library and the Hyperion hyperspectral image were spectrally matched to obtain the vegetation type and its distribution in the study area.The main conclusions are as follows:(1)Comparing the mixed spectrum library established by the two methods of adding correction and not adding correction,it is found that adding correction to the measured spectrum can greatly improve the matching accuracy.The root mean square error of more than 99% of the pixels is less than 0.1,which is significantly better than the measured spectrum The matching accuracy of the created spectral library is not added.(2)The matching accuracy of the five matching algorithms from high to low is SAM,SID_SAM,SID,SCM and SCA.(3)According to the vegetation classification map of each algorithm,the main vegetation types in the study area are mixed vegetation of Kobresia humilis and Alpine Carex and Stipa.Compared with the single vegetation mixed spectrum,the double vegetation mixed spectrum is matched.More,corresponding to a wider field area,consistent with the results of field surveys.