Screening and Artifact Detection of RFI in Sentinel-1A Time-Series Images Combining Change Detection Techniques With Structural Similarity Index

As a wideband radar system, spaceborne synthetic aperture radar (SAR) is susceptible from other high-power radiation sources, which can cause radio frequency interference (RFI) artifacts in the acquired images. Given that RFIsignificantly impacts data processing and image interpretation, screening and artifact detection of RFI have become top priorities for spaceborne SAR systems which generate massive data daily. However, current image-level methods mostly rely on the acquisition of relevant prior knowledge, which makes efficient screening of large-scale products difficult when the scenario changes. In this paper, a novel screening and detection method based on change detection techniques is proposed. Due to the time-varying feature of RFI artifacts in time-series SAR images, local areas affected by RFI can be masked through change detection operator. The screening step is performed by quantifying and analyzing the feature differences between RFI artifacts and ground-truth information in local areas. Therefore, an RFI-free background is constructed based on the results, enabling effective artifacts detection of RFIcontaining images. Experimental results of Sentinel-1A Level-2 products verify the effectiveness and robustness of the proposed method.