Change detection in repeat pass interferometric synthetic aperture radar

In this paper, a scene change statistic based on a log likelihood ratio has been considered for detecting man-made scene disturbances in repeat pass SAR interferometry. The change statistic identifies scene disturbances through changes in the scene coherence as well as changes in the scene backscatter coefficient. Theoretical analysis shows that the statistic gives significantly better detection performance than that of the commonly used sample coherence and backscatter coefficient ratio change statistics. The performance improvement has been experimentally verified using repeat pass SAR data collected with the DSTO Ingara SAR.