An advanced change detection method based on object-oriented classification of multi-band remote sensing image

An advanced change detection method based on object-oriented classification of multi-band remote sensing data is proposed in this paper. In the first step, a further classification image is obtained based on the processed segment image and the preliminary classified image. Then the change threshold is achieved by an improved adaptive recursive algorithm based on the fusion image. Finally, the change region and change category are obtained by multi-band integration change mask method. The experiments showed that this method is effective, it can improve the accuracy of change detection and determine type of change, and can be used as a robust tool for territorial resources survey.