Detecting and Mitigating Radio Frequency Interference Artifacts via Tensor Decomposition of Multi-Temporal SAR Images

Radio frequency interference (RFI) in space-based radar echo signals may affect the coherent focus imaging process, resulting in blurred scattered images or occlusion artifacts. Conventional echo domain RFI mitigation methods do not work well with image domain data. Therefore, a novel RFI mitigation method based on tensor low-rank sparse decomposition in the image domain is proposed in this paper. The tensor low-rank sparse decomposition problem can fully preserve the spatial correlation between images. A joint mathematical model of low-rank sparse tensor decomposition is established and solved to achieve the extraction and mitigation of remote sensing image interference. The results of Sentinel-lA data show that the method can extract interference artifacts and recover clear background images. The interference mitigation performance of this method is better compared with the previously proposed matrix decomposition method.