SAR super-resolution with a stochastic point spread function

Super-resolution techniques attempt to improve upon the theoretically optimum resolution of a coherent imaging system by using prior knowledge about the imaged scene. Ideal operation requires the exact point spread function (PSF) of the imaging system to be known. For airborne SAR systems, even after motion compensation, residual across-track motions of the SAR platform will cause unknown perturbations of the PSF. In this paper, the effects of such perturbations on the performance of a particular super-resolution technique are illustrated. A modification to the super-resolution technique which uses a stochastic description of the PSF is introduced as a possible means of limiting the degradations caused by PSF perturbations. The performances of the standard technique and the modified technique are compared