SAR change detection techniques and applications

Change detection, the comparison of remote sensing images from different moments in time, is an important technique in environmental earth observation and security. SAR change detection is useful when weather and light conditions are unfavourable. Five methods of SAR change detection are presented and evaluated. The performance of the first one, post-classification change detection, is strongly dependent on the accuracy of classification. Because the latter can be a problem in SAR images, this is not a proper method. The second one is CFAR detection. This method is only suitable when the changes are small compared to the resolution. The third method uses an adaptive filter and is able to deal with distributed changes as well. Besides, it can be more effective in case of small changes, but it lacks in reproducing the shape of changes. The fourth method makes use of multi-channel segmentation, a method that reproduces the shape of changes well for multi-look images, but lacks in the detection of small changes. The fifth method is a combination of CFAR detection, adaptive filtering and multi-channel segmentation. This method combines the appropriate detection of distributed and small changes, even for noisy SAR images. As a result there is a preference for the last three methods, but what method can best be used is dependent on the application too.