Image Deblurring Based on Edges Information and Markov Random Field

An image deblurring model based on edges information and improved image prior knowledge is proposed, and the edges information and Markov Random Field is used to establish a priori knowledge which is suitble for all types of picture. The fuzzy kernel of fuzzy image can be estimated according to the proposed new model, which gets a better blurring effect. Through the comparison experiment results for three images of different types of fuzzy picture show that the proposed deblurring method has a stronger universality, and obtains more clear visual effect, the Peak Signal to Noise Ratio is respectively up to 26.79 dB, 26.33 dB and 29.81 dB, and the Mean Structure Similitary Index is respectively up to 0.86, 0.88 and 0.90, comparing with the existing algorithms, the proposed algorithm is improved obviously.