城市建设用地遥感信息提取方法研究

Remote sensing information extraction of urban built-up land

本文利用CBERS-02B影像,通过分析各个地类的光谱特征,发现了长沙市城市建设用地和其他背景地物的区别,并在此基础上选取了土壤调节植被指数(SAVI)、归一化水体指数(NDWI)和比值居民地指数(RRI)作为三个指数波段,重新进行波段组合,从而减少了波段数据的冗余,最后采用最大似然法分类,提取出城市建设用地信息,其正确率达到85.6%。

Based on the CBERS-O2B image data, this paper analyzed the Spectral Characteristics of land use types, and found the difference between the urban built-up land and others. The study selected three indices, i.e. Soil Adjusted Vegetation Index （ SAV I）, normalized difference water index （NDWI） and Ratio Resident-area Index（RRI） , to represent three major index bands generated from the original muhispectral bands and dramatically decreased band correlation. Using Maximum Likelihood Classification （MLC） to retrieve urban built-up areas in the region of Changsha, the overall accuracy of them were 85.6%.