Enhanced-resolution SAR tomography using the weighted covariance fitting criterion

Abstract:
The usage of maximum-likelihood (ML) regularization to solve the synthetic aperture radar (SAR) tomography (TomoSAR) inverse problem has been successfully demonstrated in previous studies. Among their main advantages are the retrieval of resolution-enhanced tomograms from a reduced number of passes, the suppression of artifacts and the reduction of ambiguity. Yet, this approach is constrained to the assumption that the observed data has a Gaussian probability density function (pdf). Consequently, with the aim of easing this restriction, preserving the previously mentioned desired characteristics, we suggest applying the weighted covariance fitting (WCF) criterion instead. The feature enhancing capabilities of the proposed WCF-based technique are corroborated via two data sets. First, we make use of L-band airborne TomoSAR data of the German Aerospace Center (DLR), acquired by the F-SAR sensor over a forested test site in Froschham, Germany (2017). Next, we refer to L-band airborne TomoSAR data of the Jet Propulsion Laboratory (JPL) from the National Aeronautics and Space Administration (NASA), acquired by the UAVSAR system over the urban area of Munich, Germany (2015).