Image Super Resolution Reconstruction Algorithm Based on Multiple Prior Constraints

High spatial resolution is necessary for several applications such as visual inspection. However, the conflict between resolution and image distance limits the applications of image devices. In this paper, a super-resolution framework with multiple priors is proposed. Firstly, the directional generalized total variational and the non-local self-similar constraint are incorporated to enhance image texture details and smooth edge effects. Especially, an adaptive Gaussian kernel is used to better descript the non-local prior. Secondly, the proposed multi-constraint problem is solved by the alternate direction multiplier method. Generally, a large number of qualitative and quantitative results demonstrated the effectiveness and superiority of our method over traditional methods.