An Efficient Dual-Domain Attributed Scattering Center Extraction Algorithm for SAR Ground Targets

Effectively extracting electromagnetic scattering features from synthetic aperture radar (SAR) images is essential for accurate radar target interpretation and recognition. The attributed scattering center (ASC) model, which integrates physical optics and geometric theory of diffraction, offers a comprehensive representation of SAR targets. However, existing ASC extraction methods face challenges such as incomplete phase information, parameter coupling, and high computational costs, especially when dealing with complex targets or low signal-to-noise ratio (SNR) scenarios. To address these issues, this letter proposes an efficient sparse recovery algorithm for ASC extraction in SAR ground targets by combining complex image domain segmentation with phase-frequency domain sparse optimization. Specifically, the principal scattering structure regions in SAR images are preliminarily segmented, and ASC parameters are subsequently estimated by matching the ASC model in the frequency domain utilizing the orthogonal matching pursuit (OMP) algorithm. A stepwise dictionary dimension reduction strategy is adopted to accelerate parameter searching and reduce computational and storage costs. Experimental results demonstrate that the proposed method significantly enhances both speed and accuracy in ASC extraction, enabling robust reconstruction of SAR ground target images with improved target feature representation.