Application of high-resolution RS image in settlement extraction

Popularity of high resolution image makes it possible to interpret the internal land use condition within settlement. The categorization of settlement is firstly made as rural (mountain, plain) and urban (city village, living block) areas. Then object-based and generic network image processing methods are primarily used in the extraction from the merged Quickbird and SPOT high resolution image. With mathematical as well as geometrical model, indices such as building density, floor area ratio, green ratio and space ratio are extracted accordingly. Meanwhile, land use map, field survey and cadastral map are employed for the estimation and accuracy assessment. In the end, analysis and comparison are made for land use condition within settlement.