An Autofocus Method based on Maximum Image Sharpness for Fast Factorized Back-projection

Fast Factorized Back-projection (FFBP) is a new way to fix the high computation problem in Back-projection (BP) but it still keeps high-resolution and strong adaptability in synthetic aperture radar (SAR). However, Applying conventional autofocus methods to both BP and FFBP requires a special consideration. In this letter, we present a new method to solve the phase error problem. We establish a evaluation function based on maximum image sharpness in FFBP. Then we utilize Coordinate Descent (CD) optimization to transform multi-variables into a single variable. To deal with the single variable problem we introduce a secant method. The proposed method keeps the rapid imaging performance of FFBP and well solves the problems caused by phase errors, and has faster convergence speed. Using the simulation data to verify the effectiveness of the proposed algorithm.