Research On Blind Motion Image Deblurring Problem

For the blind restoration problems of blurred degraded image caused by movement, the principle of motion blur image formation, classical restoration methods, variational model, Bayesian model and the kernel estimation technology methods are studied in this dissertation. This dissertation uses the rational polynomials, which can better fit the probability distribution of natural information in the image. Meanwhile, blind restoration frame of the motion blurred image based on the mixed information is proposed. The frame can retain more details and suppress noise well through experimental verification algorithm. The main work is divided into four parts as follows:(1)Firstly, the meaning and purpose of image deblurring is introduced, the overview and development of single motion blurred image restoration technology is analyzed and the framework of this dissertation is given.(2) Image degradation model and the relative basic knowledge as well as the traditional methods of image restoration are studied, then based on this all the existing algorithms are classified into two main parts and comparison is given.(3) After studying Bayesian models, a research the motion image of blind restoration method based on Bayesian frame, and making the Hyper-Laplacian model to fit the probability distribution of natural information in the image are proposed, which can optimize the algorithm proposed by Fergus. The effectiveness of the algorithm has been validated through the experiments.(4) A blind motion deblurring algorithm frame based on mixture information is proposed. Edge extraction is applied to estimate point spread function and image edge and texture information is utilized to build the image restoration model for the blurred image restoration, then split Bregman iterative algorithm is introduced to solve the restored image. In the process of blind restoration, the image is in the multiscale restoration by Laplace pyramid, avoiding interference due to the excessive blur kernel. Because in the recovery process a great deal of edge and texture information is retained, the effect of blind image restoration is significantly enhanced by experiments.