Benefit of SqueeSAR filtering in topography reconstructions: an example over the Bucharest area with TerraSAR-X

Abstract

Advanced Synthetic Aperture Radar (SAR) processing techniques have gradually become a powerful tool for earth observation. The ubiquitous problem of noise filtering represents also a challenge in the SAR signal processing field. In case of the analysis of land regions, the identification of a general algorithm, suitable for the large-scale processing, represents an additional challenge, due to the diversity of the areas’ characteristics. More specifically, two main classes of scattering mechanisms can be defined: Persistent Scatterers (PS), coherent, point-wise targets, and Distributed Scatterers (DS), mechanisms with moderate coherence, whose contribution spreads across multiple pixels of the images. Due to their different characteristics, the joint processing of PS and DS was difficult to address, until the development of the SqueeSAR algorithm. This method, based on the joint processing of statistical homogeneous pixels, was successfully implemented in the field of SAR Interferometry, which exploits solely the phase of the focused acquisitions. The objective of this work is to evaluate the applicability of the SqueeSAR principle in the frame of SAR Tomography, a multitemporal SAR processing technique which also has the ability to separate interfering scattering mechanisms.