RFI Localization Algorithm Based on Sparse Bayesian Learning for Synthetic Aperture Radiometers

Abstract

For synthetic aperture imaging radiometer (SAIR), the effect of radio frequency interference (RFI) is affecting the quality of the brightness temperature (BT) retrieval. RFI detection and localization are essential for RFI mitigation, which contributes to the retrieved data improvement. Sparse Bayesian spatial spectrum estimation algorithm has been proposed and simulated for traditional radar DOA, and experiments shows that its performance is better than traditional compressed sensing methods. Considering the common characteristic of traditional array signal processing and SAIR, This paper proposes an RFI localization method based on sparse Bayesian learning. The undersampling information is and combine it with Expectation-Maximum (EM) algorithm for joint parameter estimation. The proposed method is validated through simulation of generated RFIs and SMOS L1a data, which shows the precise performance of the proposed method for RFI detection and localization.