Space-variant Phase Error Estimation and Correction for Automotive SAR

Phase errors deteriorate the quality of synthetic aperture radar (SAR) images drastically. In the practical automotive SAR data acquisition, the phase error varies for different regions of the observed area. It seems reasonable to consider a similar error for targets in a small region of the SAR image. In this work, we combine a generalized method of the phase gradient autofocus (PGA) and the fast factorized back-projection (FFBP) algorithm to estimate the phase error and reconstruct compensated sub-images. The proposed method is evaluated with multi-target scene simulations and practical experiments by using a 77-GHz radar system. We will show that the space-variant phase error estimator improves the image quality with a lower computational cost than conventional methods.