RFI Suppression Based on Sparse Frequency Estimation for SAR Imaging

Hongqing Liu,Dong Li
DOI: https://doi.org/10.1109/lgrs.2015.2496620
IF: 5.343
2015-01-01
IEEE Geoscience and Remote Sensing Letters
Abstract:This letter addresses the problem of synthetic aperture radar (SAR) image recovery in the presence of radio frequency interference (RFI), which degrades SAR image quality if it is not effectively suppressed. In this letter, the RFI is modeled as the superposition of multiple complex sinusoids such that the RFI suppression problem is transformed to a frequency estimation problem. To accurately estimate the amplitudes of the sinusoids and their corresponding frequencies in the case of a low number of range samples, the frequency sparsity in the frequency domain is successfully exploited. From the estimated amplitudes and frequencies, the RFI can be reconstructed and then used for suppression. To recover the signal of interest (SOI) and by utilizing the estimated RFI, a joint estimation is derived to simultaneously perform the RFI suppression and the SOI recovery. This joint approach can effectively suppress the RFI even if it overlaps with the SOI in both the time and frequency domains. The common threshold decision approach is not required for our joint estimation to reduce the RFI. Simulation results and real-world experiments are presented to demonstrate the superior performance of the joint estimation algorithm.
What problem does this paper attempt to address?