Urban ecological land extraction from Chinese Gaofen-1 data using object-oriented classification techniques

The urbanization process changed the urban ecological land and consequently affected the quality of urban residents´ environment, and it was very important to obtain urban ecological land cover information. In this paper, an object-oriented method was proposed to extract urban ecological land cover from the multiple-channel images acquired by Chinese Gaofen-1 (GF-1) satellite. Taking Beijing City as an example, five ecological land covers, including water, vegetation, road, building land and bare land, were classified using new classification rules based on the spectral, geometry and texture information in the GF-1 image. The result showed that the urban land covers were accurately identified and its validation accuracy was up to 90%.