Change Detection in Multitemporal SAR Images: A Multiresolution Analysis and Hypothesis Testing Approach

Synthetic Aperture Radar (SAR) is an effective tool for detecting environmental changes over time. Its ability to provide images of high-resolution in any kind of weather conditions and its independence of daylight make it a valuable asset in monitoring and management of environmental changes. This paper presents a new method for detecting changes in multitemporal SAR images without the use of speckle filtering. The approach consists of: (1) creating a multiresolution set from the ratio image through the ratio of each scale of the two intensity SAR images after being processed with the stationary wavelet transform (SWT); (2) conducting hypothesis testing on the multiresolution ratio image set; and (3) combining the results at the optimal reliable scale to produce the resultant map of change. The effectiveness of this method was demonstrated through experiments on semi-simulated RADARSAT-2 satellite SAR images in Algiers region.