A Spectral Mixture Analysis and Mapping of Impervious Surfaces in Built-up Land of Fuzhou City
福州城区不透水面的光谱混合分析与识别制图

As a major component of the Ridd's vegetation-impervious surface-soil(V-I-S) model,urban impervious surface is important in monitoring urban expansion and understanding impacts from human activities on the environment. By means of image processing technique,the impervious surface information can be extracted from remote sensing imagery.In this paper,a minimum noise fraction(MNF) transformation has been applied to a Landsat 7 Enhanced Thematic Mapper Plus(ETM ) sub-scene of Fuzhou city,southeast China.The first three components produced from the MNF transformation were selected.Through a linear spectral mixture model,four endmembers,i.e.,low albedo,high albedo,vegetation,and soil were identified to represent the heterogeneous urban land cover types by combining low and high albedo components.Finally the impervious surface fraction was estimated.The result indicates that the increase in urban impervious surface has a negative impact on the urban ecosystem.