Automatic Detection Technique of Residential Area Change Based on Topological Analysis

Based on topological analysis, an automatic detection method of residential area change is presented in this paper. Firstly, we demonstrated the importance of extracting residential area in remote sensing images accurately and automatically. Then, it is found that the old vector of the same region even at different time is useful to obtain the train samples. By applying old vector, we classified the remote sensing image, and extracted the residential area automatically. Consequently, topological analysis is applied to detect six different types of change automatically. Experimental results of residential areas detection on SPOT images show the merit and accuracy of the proposed method.