Automatic Quality Assessment for Interferogram SAR Stacks

We propose a robust and reasonably fast approach to provide phase estimation and stable target (Persistent Scatterers, PS) detection from a set of repeated pass interferometric acquisitions. The major assumption is that, locally, the PS phase model is decomposed into the separable product of the intrinsic target phase times the optical path, that is common to all the targets in the estimation window. The Singular Value Decomposition (SVD) is exploited to get at one time a fast estimate of the local phase and the detection of stable targets. The estimated phases are then exploited to measure the PS coherence over the whole image or in image blocks. Being sensible to acquisition and processing artifacts, the PS coherence is a valuable tool for automatic image quality assessment. Results are shown with real datasets in both single and full polarization.