Cartesian based FFBP algorithm for circular SAR using NUFFT interpolation

Circular SAR is able to achieve omni-directional observation and high-resolution imaging of targets. However, the traditional frequency-domain based imaging algorithm is not suitable for complicated curve trajectory. Moreover the time domain based back-projection (BP) algorithm is applicable but time consuming. Fast factorized back-projection (FFBP) algorithm based on aperture decomposition and image fusion can balance computational efficiency and accuracy. In this paper, we proposed a modified FFBP algorithm for circular SAR imaging. The principal improvement is the usage of Cartesian coordinate imaging and nonuniform fast Fourier transform (NUFFT) interpolation. First, sub-aperture BP imaging is implemented on local Cartesian coordinate system. Then azimuth bandwidth is compressed with a spatial variant phase function to reduce the sampling rate. Next the NUFFT interpolation method is applied during sub-images fusion to further improve the efficiency of the algorithm. Finally, through simulation and real data experiments, the correctness and accuracy of the algorithm is verified.