Wideband interference mitigation for synthetic aperture radar based on the variational Bayesian method

Wideband interference (WBI) seriously detracts from synthetic aperture radar (SAR) imaging quality and hinders the subsequent interpretation of images. Meanwhile, the large bandwidth and complex modulation of WBI necessitates the development of mitigation algorithms with strong adaptability and robustness that are lacking in existing algorithms based on filtering and model analysis. The present work addresses this issue by proposing a WBI mitigation algorithm based on variational Bayesian inference (VBI). Firstly, the WBI-contaminated echo signal is identified in the time–frequency (TF) domain through an adaptive statistical detection method. Then, a low-rank matrix factorization is formulated according to the low-rank characteristics of the WBI and the Laplace distribution assumption of the target echo signal in the TF domain. Finally, the WBI component is accurately reconstructed using a mean-field VBI method, and then eliminated from the original SAR echo signal to recover the target echo signal. The effectiveness and robustness of the proposed WBI mitigation algorithm are demonstrated based on its application to simulated and measured SAR data.