Image Restoration Algorithm For Suppressing Ringing Effect

Spatial remote sensing image restoration process is the inverse process of degradation of image.In the image degradation model,the parameters that fuzzy kernel and random noise are unknown.Parameters estimate errors which will lead to the restoration of image degradation,the impact the judgment of the human eye on the visual effect.In image restoration,a phenomenon particularly serious influence the image called the ringing effect.Its influence on image quality is serious,mainly for the restoration of the original image in the original clear image does not exist in the dark and light.Based on the background of remote sensing image,suppress ringing effect after deblurring.According to the ringing effect appeared in the position of the image which can be divided into the image boundary ringing effect and edge ringing effect.The image boundary ringing effect is mainly in the frequency domain which can not meet the periodicity of the discrete Fourier transform and fast image gradient.The image edge ringing effect is due to degradation of model parameter estimating error.And it is shown that the probability characteristics of image statistics obey the Gibbs distribution.According to the boundary ringing effect of image,smoothing the gradient of the fuzzy image and the fuzzy image is used to satisfy the periodic extension by itself.Thus discrete Fourier transform to meet periodically and the boundary ringing effect does not produce boundary truncation phenomenonFor the image edge ringing effect,according to the theory of Bias,the problem which design the maximum a posteriori probability is established,and the problem is transformed into the solution of the minimum energy problem.Iterative calculation of a priori fuzzy kernel and remote sensing image.And fitting the heavy tailed distribution of the natural image with Gauss distribution This algorithm is used to calculate the fuzzy kernel,and the iterative method is used to calculate the interaction between the fuzzy kernel estimation and the restored image,and the method is gradually close to the true value.Finally smooth gradient boundary algorithm and maximum posterior probability restoration algorithm combining which design oriented ringing suppression of image restoration algorithm.The algorithm in remote sensing image blur as the basis,to suppress the ringing effect,from the experimental data and visual effects show that,the algorithm of blurred remote sensing image restoration effect is good,the algorithm is feasible.