Extraction Information of Impervious Surface Area Based on Improved NDBI

On the basis of the spectral analysis for kinds of surface features in Landsat ETM+images, this paper proposed an improved NDBI algorithm. Firstly, the mechanism of remote sensing for impervious surface was analyzed. Secondly, the difference of impervious surface and the other land-use types were discussed. Thirdly, the structure feature among spectrum of impervious surface and the other landuse types was analyzed. Vegetation could be extracted by NDVI with suitable threshold. The results prove that the proposed algorithm has high accuracy and believable results. Compared with the normal NDBI algorithm, the proposed algorithm effectively separates the impervious surface from the sparse vegetation and bare land. It greatly improves the accuracy of the extraction of impervious surface information and has the advantage of automatic and efficient.