Research on the updating method of land-use map based on object-oriented analysis

Based on remote sensing (RS), the updating method of land-use map is becoming one of the most important study direction in land use and cover change (LUCC) field over the world. This study have broken through traditional land use change detection methods based on two remote sensing images and matched the old land use vector data to the remote sensing image of new period, where making use of the vector border and classificatory attribute of the map unit in the land use map. Extracting the change part of the image through the filtering method of two thresholds, then the homogenous objects can be got through the multi-scale imaation, and the object-oriented classification has been done onge segment the image. Finally through the alternant edit with the land-use vector data, the updated land-use map can be got. The accuracy assessment showed that the result is satisfying and better than the traditional detection method. It realized the updating of the land-use data commendably and provided a new method for the updating the land-use map.