A fast autofocus algorithm for synthetic aperture radar processing

High-resolution synthetic aperture radar (SAR) imaging requires the motion of the radar platform to be known very accurately. Otherwise, phase errors are induced in the processing of the raw SAR data, and bad focusing results. In particular, a constant error in the measured along-track velocity or the cross-track acceleration leads to a phase error that varies quadratically over the synthetic aperture. The process of estimating this quadratic phase error directly from the radar data is termed autofocus. A novel autofocus algorithm with a computational complexity which is at least an order of magnitude lower than that of other algorithms providing comparable accuracies is presented. The algorithm has been tested on data from the Danish Airborne SAR, and the performance is compared with that of the traditional map drift algorithm