Research on Extracting Residential Information from ETM+Images - a Case Study of Jiaxing City

摘要

Extracting residential area from remote sensing image is an important issue in remote sensing mapping, disaster assessment, city planning and other field. In this paper, the existing residential area extraction method are analyzed at first, including the method of threshold structure in the spectral, combination knowledge discovery and spectral structure, NDBI, NBI and IBI; Then, the differences of Spectral characteristic and how to distinguish residential area and other land use type are discussed, the extraction model is build based on the above research. Taking Jiaxing City, Zhejiang Province as an research area, residential area are extracted from ETM+ image use NDBI index and Spectral structure difference. At last, precision analysis indicate location accuracy is 85%, area accuracy is 88.8%, the result show this method can improve the divisibility of residential area, bare land and road. The extraction result can provide data for remote sensing mapping, disaster assessment, city planning, etc.