Fusion of features in multi-temporal SAR imagery to detect changes in urban areas

The Dempster-Shafer (D-S) algorithm is improved to fuse different features of multi-temporal Synthetic Aperture Radar (SAR) images to detect changes in urban areas. Firstly, D-S theory is developed by not only considering the certainty of the sources, but also considering the average support of the sources to different subsets in the frame of discernment, in the process of evidence combination. In this way, it can assign conflict information in deferent sources reasonably and present a more reliable combination result. Secondly, amplitude ratio and distance between probability density distribution functions in Pearson graph of different temporal SAR images are extracted to present the change features in different scales. Lastly, the improved D-S algorithm is applied to fuse the two different features to detect change information of SAR images. An example of the Shanghai Lujiazui area using European Remote Sensing 2 (ERS-2) SAR images demonstrates the accuracy of the improved fusion algorithm.