Object-oriented Per-parcel Land Use Change Detection Integrating Gis and Remote Sensing

Urban is more sensitive to the change of land use/cover. The up-to-date information of land use/cover is very important for urban decision-making and sustainable development. Very high resolution satellite data are an alternative to monitor, update cartographic and geographic databases. In this paper, an object-oriented pre-parcel land use change detection approach integrating GIS and remote sensing is introduced. The change detection unit is defined by existing objects in a GIS databases, different image processing program are adopted based on the corresponding objects attributes. The approach mainly consists of two phases: first, land cover objects are identified using object-oriented image classification approach. Next, a rule-based system is set up to judge every parcel one by one whether or not change happened comparing to existing GIS. The result is marked as change, unchange and unclear, the unclear part will be decided further by interactive checking. In order to verify the application validity of the presented methodology, the part of Shanghai in China with the support of QuickBird data and GIS is tested, a satisfactory result is attained.