RFI mitigation for SAR based on compressed sensing and morphological component analysis

Abstract

High-resolution synthetic aperture radar (SAR) which operates with a large bandwidth is subjected to the impact from various kinds of radio frequency interference (RFI) in complex electromagnetic environment. RFI mitigation from the contaminated SAR signal is an appealing issue in the field of radar signal processing. Conventional techniques of interference mitigation such as notch filtering may lead to side-lobe effects and degrade the quality of imaging. This paper presents a novel method of RFI mitigation for SAR base on compressed sensing and morphological component analysis (MCA). First, dictionary of target echo and RFI is respectively constructed with their individual characteristic in sparsity domain. The sparsity of target echo and RFI are then demonstrated to be mutual exclusive in each other's sparse dictionary, the condition of which satisfies the prerequisite of utilizing MCA. Finally, RFI is separated from the compressed measurements and the clean signal is reconstructed with the stepwise orthogonal matching pursuit (OMP) algorithm. Experimental results of simulated data demonstrate the effectiveness and advantage of our proposed method.