Improved iteration phase gradient autofocus CFBP for portable near-field SAR real-time imaging

For flexible and rapid stealth equipment radar cross section (RCS) measurement, a portable near-field Synthetic Aperture Radar (PNF-SAR) system is designed. However, the SAR image tends to defocus and be unstable since motion errors are introduced by mechanical inaccuracies and environmental disturbances. The paper proposed an improved iteration phase gradient Cartesian Factorized Back Projection (CFBP) algorithm, combining the advantages of the CFBP framework with image-based autofocusing techniques and constructed a phase error model based on a contrast maximization criterion, calculated by iteration conjugate gradient solver with an Armijo line search algorithm. To reduce computational load, the algorithm restricts optimization to local regions around strong scatterers, avoiding global operations while maintaining imaging accuracy. Additionally, least-squares fitting is applied to smooth the estimated phase errors, further improving robustness. Simulation and analysis show that this method achieves a good balance between imaging quality, computational efficiency and practicality in near-field SAR imaging.