Adaptive Multi-Objective Pulse Compression for SAR Imaging

For synthetic aperture radar (SAR) imaging, the classical matched filter produces high sidelobes that bury weak targets, while windowing sacrifices resolution. Existing mismatched filter designs and adaptive pulse compression methods tend to concentrate on local sidelobe suppression or adopt restricted optimization criteria, resulting in insufficient global control and limited practical effectiveness. To address these issues, a cascaded framework that combines mainlobe–fidelity dual–norm iterative reweighted suppression (MDIRS) with local normalized sharpening (LNS) is proposed. The proposed MDIRS–LNS chain achieves high dynamic range, enhanced weak–target visibility, and a resolution comparable to the matched filter, while attaining a balance between sidelobe suppression depth and resolution retention compared with conventional linear mismatched filters and windowed matched filters. Experiments on simulated and real satellite SAR imagery data are provided to validate the effectiveness of the proposed method.