A Virtual Adaptive Beamforming Approach for Feature Enhanced SAR Tomography

Abstract

Synthetic aperture radar (SAR) tomography (TomoSAR) is a remote sensing technique that allows for the 3-D representation of the illuminated areas, recovering the vertical distribution of the backscattered power at each range-azimuth position. In this context, and with the aim of retrieving feature-enhanced tomograms, this paper addresses a new multi-stage iterative method that operates robustly in real-world TomoSAR operating scenarios with irregularly distributed acquisition constellations and only few available looks. The addressed approach alleviates the drawbacks of the conventional matched spatial filter (MSF) technique, which retrieves high ambiguity levels when irregular sampling is considered. Also, it alleviates the drawbacks of the commonly used Capon beamforming technique, which results to be inapplicable when the involved data covariance (structure) matrices are rank deficient. The addressed novel method combines the descriptive experiment design regularization (DEDR) framework for enhanced image reconstruction, with the wavelet domain thresholding (WDT)-based sparsity promoting refinement in the wavelet transform (WT) domain. The capabilities of the addressed WDT-refined virtual adaptive beamforming (VAB) approach, which we refer to as WAVAB, are corroborated via processing P-band airborne TomoSAR data of the German Aerospace Center (DLR), acquired by the E-SAR system over the test site located at the Vindeln municipality, northern Sweden, in 2008.