Superresolution and edge-preserving reconstruction of complex-valued synthetic aperture radar images

We propose a regularization-based method for the complex-valued synthetic aperture radar (SAR) image formation problem. The method can produce images with higher resolution than that supported by the measured data, as well as images with reduced variability of reflectivity magnitudes within homogeneous regions and preserved region boundaries. This is achieved by the inclusion of prior information in the regularized data inversion process through nonquadratic potential functions. For this task, we demonstrate the use of a variety of potential functions in our framework. The technique effectively deals with the random-phase nature of the underlying SAR reflectivities. For an efficient numerical solution, we extend half-quadratic regularization methods to this complex-valued problem. We demonstrate the performance of the method on real SAR data