Sparse Model and Sparse Recovery with Ultra-wideband SAR Applications

This paper presents a simple yet very effective timedomain sparse representation and the associated sparse recovery techniques that can robustly capture raw ultra-wideband (UWB) synthetic aperture radar (SAR) records. Unlike previous approaches in compressed sensing for SAR, we take advantage of the sparsity and the correlation directly in the raw received pulses before even attempting image formation. Recovery results from real-world data collected by ARL’s ultra-wideband (UWB) SAR system illustrate the robustness as well as effectiveness of our proposed framework on three diverse applications: subNyquist sampling of raw SAR data records, recovery of missing spectral information in multiple frequency bands, and extraction as well as suppression of radio frequency interference (RFI) signals from SAR data records. Keywords-UWB, SAR, compressed sensing (CS), sparse recovery, sparse noise, sub-Nyquist, missing spectrum, RFI