A robust multiframe super-resolution algorithm based on half-quadratic estimation with modified BTV regularization

Multiframe image super-resolution is a technique to reconstruct a high-resolution image by fusing a sequence of low-resolution images of the same scene. In this paper, we propose a new multiframe image super-resolution algorithm built on the regularization framework. The objective functional to be minimized for the regularization framework consists of a fidelity term and a regularization term. A new adaptive norm combining the advantages of traditional L1 and L2 norms is used in both terms. The fidelity term is then formed by an adaptive strategy depending on the accuracies of the estimated low-resolution image observation models. This strategy serves to adaptively weight low-resolution images according to their reliability and can add robustness in practical implementation of super-resolution. The proposed regularization term can preserve sharp edges well without producing visual artifacts. Our experimental results using both synthetic and real data show the performance improvement of the proposed algorithm over other methods.