Wide-Band Interference Mitigation for SAR Based on Generative Adversarial Network

The presence of wide-band interference (WBI) seriously affects the imaging quality of synthetic aperture radar (SAR) as well as the target scattering characteristics, which reduces the accuracy of subsequent SAR image interpretation. This paper proposes a WBI interference mitigation algorithm for SAR based on generative adversarial network (IM-GAN), which can extract the time-frequency features of WBI in time-frequency domain to achieve low-loss reconstruction of the target echo signal. Firstly, the short-time Fourier transform (STFT) is used to transform the radar echo signal into the time-frequency domain. Then, the GAN is utilized to reconstruct the target signal by extracting the characteristic differences between WBI and target echoes. Finally, the time-frequency spectrogram of SAR echo signal after interference mitigation is transformed to the time domain using inverse short-time Fourier transform (ISTFT). The effectiveness of IM-GAN is verified by the WBI mitigation experiments of the measured dataset in TOPSAR mode recorded by the Sentinel-1B, and its superiority is demonstrated by the comparison results of different SAR interference mitigation methods.