Enhanced RBSF for 2D-ISRJ Detection, Localization and Mitigation in SAR Raw Data
Abstract
Synthetic aperture radar (SAR) plays a pivotal role in remote sensing owing to its all-day and all-weather imaging capability. However, in complex electromagnetic environments, SAR systems are vulnerable to severe jamming threats. Among the various jamming types, 2-D intermittent sampling repeater jamming (2D-ISRJ) is particularly representative, as its high coherence with the radar signal generates dense false targets that severely degrade SAR image quality. As an advanced subspace-based technique, robust block subspace filtering (RBSF) exhibits strong robustness in jamming mitigation against strong-scatterer backgrounds. Nevertheless, RBSF identifies and protects strong scatterers prior to mitigation. Owing to the high coherence between 2D-ISRJ and strong scatterers, the jamming is falsely classified as a strong scatterer and thus shielded from mitigation, leading to severe performance degradation. To address the above limitations, this article integrates 2-D fine masking (FM) and time–frequency (TF) weighted fusion, yielding an advanced method termed fine-mask robust block subspace filtering (FM-RBSF). Specifically, block subspace filtering is first employed to generate a reference signal in which the jamming and the strong scatterers are simultaneously stripped. Subsequently, time–frequency information entropy is jointly exploited with multidimensional morphological features to construct a 2-D fine localization mask of the 2D-ISRJ in the TF domain, thereby decoupling the jamming features from those of the strong scatterers. Finally, guided by the fine mask, the purified signal components and the original echo are integrated through TF-domain weighted fusion, enabling the reconstruction of the strong scatterers while suppressing the jamming. Experiments conducted under various jamming-to-signal ratios verify the effectiveness and robustness of the proposed method.