Blind Super-Resolution Image Reconstruction Based on Weighted POCS

Abstract

In the paper a weighted Projection onto to convex sets (POCS) algorithm for image super-resolution reconstruction is proposed, which using POCS scheme to reconstruct image without knowing the degradation model and related parameters. In the proposed algorithm, adaptive block matching is utilized to estimate translational motion among various images firstly, and then APRX algorithm is adopted to estimate the point spread function (PSF) and the ideal high resolution image. Finally based on the traditional POCS algorithm, the paper introduced the prior information about low resolution image definition into POCS to calculate the residual and control the threshold by weighted parameter of image relative definition, and realized the dynamic threshold selection in the whole progress of reconstruction. Results of experiments show that the algorithm in the objective evaluation and visual effects achieves better reconstruction results, which confirm the validity and robustness of the proposed algorithm.