Pulse RFI Mitigation for SAR Data by Integrating Hankel Structure and Truncated Nuclear Norm Regularization

Radio frequency interference (RFI) caused by spectrum allocation issues has long affected the signal processing and image interpretation of synthetic aperture radar (SAR). The pulsed version of RFI, known as PRFI, creates artifacts that are easier to suppress targets of interest because of its high power and wide bandwidth. A class of time-domain notch filtering-based approaches has been developed to tackle these sophisticated PRFIs. However, the proportion of PRFI in SAR data will increase with the gradual deterioration of the electromagnetic environment, and these methods will inevitably lose too many useful signals, resulting in severely limited mitigation performance. In order to overcome the above problems, in this article, an approach, based on the Hankel structure and a low-rank approximation model regularized by truncated nuclear norm (TNN), is developed to simultaneously mitigate powerful PRFI and protect useful signals. In the proposed approach, the Hankel-structured SAR pulses are linearly extending the low-rank property of PRFI. The TNN regularization is then employed to obtain a low-rank approximation that outperforms the traditional nuclear norm, thereby separating and suppressing PRFI. Simulated SAR data as well as measured SAR data acquired by ESA Sentinel-1A are applied to demonstrate the effectiveness and superiority of the proposed approach.