A fast augmented Lagrangian approach for compressed SAR imaging

In this paper we present an accelerated Augmented Lagrangian Method for the solution of constrained convex optimization problems in the Basis Pursuit De-Noising (BPDN) form. The technique relies on on Augmented Lagrangian Methods (ALMs), particularly the Alternating Direction Method of Multipliers (ADMM). Here, we present an application of the Constrained Split Augmented Lagrangian Shrinkage Algorithm (C-SALSA) to SAR imaging, while introducing a method to handle complex SAR imagery in the constrained Total Variation Minimization formulation. In addition, we apply acceleration schemes to C-SALSA to obtain faster convergence of the method; such as used in Fast ADMM methods proposed by Goldstein et al., in the Fast Iterative Shrinkage-Thresholding Algorithm (FISTA) proposed by Beck and Teboulle, and in NESTA proposed by Becker et al. We present examples to illustrate the effectiveness of Accelerated C-SALSA in the context of SAR imaging.