Study of Land Surface Composition of Wuhan Citybased on Linear Spectral Mixture Analysis

Rapid urban growth makes monitoring and analysis of urban surface compositions become the most significant research hotpot.Moreover,satellite remote sensing data is the major source of collecting and surveying geographic information quickly and conveniently.The Linear Spectral Mixture analysis Model (LSMM) has been reported in other literature and has been proven to be the intuitional and effective approach for describing urban surface compositions.In this paper,we try to get the real surface compositions of Wuhan City through the combination of the traditional 3-endmember method for masking water off and 4-endmember method with water based on the LSMM method using Landsat 5 thematic mapper image acquired on September 6,2009.The results show that the 4-endmember (water,vegetation,low albedo,high albedo) method is better than the 3-endmember method,because masking water off can cause major errors in water boundary regions.