SAR Imaging of Swing Ship Based on Physical Struecture-guided Segmentation

Synthetic Aperture Radar (SAR) has been widely employed in ship detection due to its ability to provide imaging regardless of weather conditions or time of day. However, external factors such as ocean waves and wind often cause noncooperative ship movements, resulting in spatial variance and severe image defocusing. Traditional Phase Gradient Autofocus (PGA) algorithms struggle to meet the requirements for high-precision imaging and recognition in such scenarios. To address this issue, this paper proposes a novel Segmented Phase Gradient Autofocus (S-PGA) algorithm to address the challenges of ship defocusing caused by noncooperative motion. The approach begins with the Range-Doppler (RD) algorithm for coarse imaging and ship region extraction. Subsequently, the ship image is divided into subregions, and the improved PGA algorithm is applied independently to each subregion for focusing. Finally, the subregion results are seamlessly stitched to produce a fully focused image. Experimental results on both simulated and real data validate that the proposed method substantially improves the focusing quality and recognition accuracy of SAR images for ships.