Analysis and Mitigation of RF Interferences in X-band SAR Data

In recent years, terrestrial Radio Frequency Interference (RFI) from airports, harbors or air surveillance systems, as well as spaceborne RFI from other satellites, has increasingly affected spaceborne Synthetic Aperture Radar (SAR) data. This paper analyzes X-band SAR data acquired by the German SAR satellites TerraSAR-X and TanDEM-X. Different approaches to directly detect and mitigate RFI in the raw data are proposed and assessed. In order to automate interference detection, machine learning methods, namely the Isolation Forest (iForest), Autoencoder, and Variational Autoencoder (VAE) are evaluated and compared in terms of their detection performance. The Autoencoder demonstrates a superior performance compared to the other methods. It is further benchmarked against the conventional Cell Averaging-Constant False Alarm Rate (CA-CFAR) detector and a basic Signal-to-Interference Ratio (SIR) filter. Results regarding detection performance and interference mitigation show the strong potential of the Autoencoder. Finally, noise pulses extracted from global TerraSAR-X acquisitions are analyzed in terms of RFI bandwidth, RFI power, and spectral RFI source density. ... mehrThis enables a global mapping of RFI occurrences in X-band data and demonstrates the effectiveness of the proposed Autoencoder framework.