Super-Resolution SAR Imaging via Compressive Sensing

This letter proposes a compressive sensing (CS)-based super-resolution (SR) synthetic aperture radar (SAR) imaging framework. The proposed method first estimates dominant scatterer support from SAR phase-history data using a two-dimensional modified orthogonal matching pursuit algorithm. The scattered signal is then reconstructed with a parametric scattering model motivated by the geometrical theory of diffraction, and the effective bandwidth is synthetically expanded in both range and azimuth. A two-dimensional Fourier transform with the modified bandwidths is finally applied to generate a super-resolved SAR image without additional measurements. Unlike conventional spectral-sharpening approaches, the proposed method performs SR through signal-domain reconstruction, enabling flexible resolution enhancement. In addition, the sparse parametric representation is expected to provide improved robustness to background noise. Experimental results demonstrate that the proposed method improves the spatial resolution of SAR images while preserving dominant target scattering characteristics.