Squint GEO SAR Imaging Based on Nonuniform Fast Fourier Transform and Modified Range-Doppler Azimuth Chirp Scaling

Geosynchronous synthetic aperture radar (GEO SAR) operating in the squint mode has the advantages of wide coverage and short revisit, which can improve the ability of emergency response and continuous observation compared with broadside mode. Due to the sensitivity of GEO SAR to the squint angle, a large amount of data redundancy and severe 2-D spatial variation problems are encountered in imaging processing. By utilizing the variable pulse repetition interval (VPRI) technique, the data reception window is adapted to accommodate changes in range cell walk, thereby eliminating redundant data. The current frequency-domain imaging algorithm is not suitable for nonuniformly sampled data. This article proposes a method based on nonuniform fast Fourier transform (NUFFT) and modified range-Doppler azimuth chirp scaling (RD-ACS) algorithm for processing squint GEO SAR data. Azimuth NUFFT can obtain data with equidistant Doppler frequencies. A 2-D frequency spectrum of reference points within the scene is used for bulk phase compensation. The slant range is represented by a high-order polynomial containing 2-D spatial variations. During range cell migration (RCM) correction and azimuth compression process, the varying coefficients of slant range with range are used to replace the corresponding reference point coefficients for range space-variant compensation. The azimuth spatial variation is addressed by the CS method, and the phase introduced by the scaling equation is compensated and the azimuth compression is completed in the RD domain. The computational complexity of the algorithm is discussed. Finally, the correctness of the proposed algorithm is verified by simulation experiments.