POCS-Based Super-Resolution Image Reconstruction Using Local Gradient Constraint

Abstract

In order to suppress and reduce the ringing artifacts of the super resolution reconstruction results, an improved POCSbased reconstruction algorithm is presented in this paper. Local gradient consistency constraint is adopted to reduce the estimation error of traditional results computed with gray value consistency. The fi xed threshold of convex sets constraint is also replaced by an adaptive version. Finally, the advanced algorithm is implemented with a relax factor of projection operation. The accuracy and effectiveness of the proposed algorithm is verified by some experiments.