A Spaceborne SAR RFI Suppression Method Based on an Adaptive Subband-Forming Technology

Abstract

Radio frequency interference (RFI) severely degrades the quality of spaceborne synthetic aperture radar (SAR) images. The most current methods usually mitigate interference in only a single domain, such as notch filtering (NF) in the frequency domain, beamforming (BF) in the spatial domain, or low-rank and sparse methods in the energy domain. Inspired by BF techniques, this article designs an adaptive subband-forming technology (SFT) model that combines both the frequency and energy domains (EDs) for interference suppression. The proposed technology consists of two processing modes: the model-solving SFT (MS-SFT) mode and the weight-application SFT (WA-SFT) mode. By introducing spectrum division into a low-rank optimization model, the model can be solved by the alternating direction method of multipliers (ADMM) with closed-form updates, and then a signal component, an RFI component, and a set of optimal subband weights can be obtained from the raw polluted SAR data. After solving the proposed model, the optimal subband weights can be applied directly to reconstruct the clean scenes and mitigate the homologous interference (HI) with similar spectral characteristics in other regions without iteration, by using the WA-SFT algorithm. Experiments on Sentinel-1 C-band data and LuTan-1 L-band data demonstrate the effectiveness of the proposed MS-SFT and WA-SFT modes.