Adaptive change detection of remote sensing images based on NDR-CVA
For the two key issues of difference image constructing and threshold segmentation in remote sensing image adaptive change detection,a fusion method of change vector analysis based neighborhood difference and ratio is introduced firstly in this paper to realize the wide use of the fusion method in remote sensing image,and effectively extract the change information.Then,the improved fast 2D-OTSU algorithm is used to segment the threshold,and to improve the automation of change detection process.Finally,it is proved that the method is an effective way in remote sensing image adaptive change detection through the experimental comparison and precision evaluation by using the Landsat 7 and Landsat 5 remote sensing images.