Research For Super-resolution Image Reconstruction Algorithm Based On MRF Prior Model

Super-resolution image reconstruction (SRR) refers to a resolution enhancement technology that extracting image information from several low-resolution (LR) images to eliminate various noise and blur which affect the quality of the image, rebuilding a higher resolution image with more clearly quality. While collecting and processing the digital images, there are many noise and degradation factors which are unpredictable, this makes the super-resolution reconstruction become to a sick reverse process. And an image model is an additional prior knowledge or constraints attached, which can make the image regularization. Therefore, the selection of the image prior model is very important to the quality of the reconstructed result.In this paper, we choose the selection of the image's prior model to be the main line, focus on the reconstruction algorithm which based on the Markov random field (MRF) prior model, have a depth analysis and experimental comparison to various kinds of reconstruction algorithm which based on the MRF prior model. Against the advantages and disadvantages of the existing MRF prior model, this paper proposed a MRF prior model which based on a new potential function, it can preserve more image edge details while suppress image noise in reconstruction. Both the peak signal to noise ratio (PSNR) and structural similarity index (SSIM) of the reconstructed image verify the validity of the new algorithm.The generalized MRF model can increase the strength of the image edge information by using the bilateral filter, to have a better result in preserving the edge feature while smooth the noise of the image. In this paper extend the new MRF model we proposed to the generalized model to improve it by bilateral filter, the experimental results show that this improvement can get better reconstruction result.