Automatic Focus and Registration of Synthetic Aperture Radar Images

A tomographic model explains why aperture phase error produces cross-range blur. New algorithms estimate phase error then correct it to remove blur from a synthetic aperture radar (SAR) image. A new sinusoidal map drift autofocus (SMDA) algorithm removes high-order (> 5) phase error better than a conventional polynomial map drift autofocus (PMDA) algorithm. A new least squares autofocus (LSA) algorithm focuses an image in fewer iterations than required by a conventional phase gradient autofocus (PGA) algorithm. A new algorithm aligns, or registers, a complex-valued SAR image with a reference SAR image. Tests indicate that the algorithm can achieve simultaneously RMS errors of 0.01^circ, 0.03 pixel, and 0.01% in rotation, translation, and scale, respectively.