High resolution differential interferometric stacking via adaptive spatial phase filtering

We present a new adaptive spatial filtering algorithm for accurate estimation of differential interferometric phase and coherence, to better support deformation monitoring techniques such as SB AS. The objective is high resolution interferometric phase estimation even in non-urban areas, where there is high phase noise. The typical method to enhance interferograms is a complex boxcar (rectangular) multilooking, leading to a low spatial resolution. This is a disadvantage, especially, with the launch of advanced SAR satellites such as TerraSAR-X. We propose a technique which provides improved interferograms without loosing the high geometrical resolution. Also, the coherence of the interferograms is estimated, which is important for various applications e.g. guiding phase unwrapping algorithms, selection of pixels for deformation monitoring etc. Our technique is based on identifying statistically homogenous pixels in a neighbourhood, and using them for phase flattening and finally, phase filtering via adaptive complex averaging (multilooking). We demonstrate our technique on TerraSAR-X data of Germany´s Bierwang gas reservoir.