An N2.5 back-projection algorithm for SAR imaging

In this paper, a modified fast back projection (FBP) algorithm is proposed for Synthetic Aperture Radar (SAR) imaging. The azimuth compression of the proposed FBP is divided into sub-image process and finial image process. The secondary phase correction is used to reduce the error cause by approximation before the finial image process. The interlace sub-image grid method makes image quality uniform for all the pixel in the finial image. The applicability limitations and the sub-image two dimensions over sample method are presented. The computational complexity of the proposed FBP is O(N2.5). Simulation on GPU is given to verify the proposed method efficacious.