A Change Detection Technique for Repeat Pass Interferometric SAR

Abstract—In the coherent change detection application of repeat pass SAR interferometry (InSAR) the scene coherence is used to detect areas of man-made scene disturbance. The sensitivity of this technique however is dependent on the extent of decorrelation across the scene that arises due to processing aberrations, system noise and environmental effects. Formulating the detection problem as a hypothesis testing problem leads to a new log likelihood change statistic with significantly improved detection performance [1]. The statistic however, requires accurate estimates of the unknown hypothesis parameters to be obtained from the interferometric image pair. This paper extends work by the authors [1] by examining the loss in detection performance that arises due to errors in the estimates of the scene partial coherence and interferometric phase required in the formulation of the log likelihood change statistic. The technique is applied to a repeat pass InSAR image pair collected using the DSTO Ingara X-band SAR where these parameters must be estimated adaptively over the scene.