Parallel processing of the fast decimation-in-image back-projection algorithm for SAR

Abstract

Fast back-projection algorithms provide substantial speedup when compared with the standard back-projection algorithm. However in many potential near-field applications of synthetic aperture radar, further speedup is still required in order to make the application operationally feasible. In this paper we investigate the application of multi-core central processing units and graphic processing units, which are now standard on most scientific workstations, to further speed up a very recently proposed fast back-projection algorithm (the fast decimation-in-image back-projection algorithm).