An Iterative Blind Deconvolution Image Restoration Algorithm Based on Adaptive Selection of Regularization Parameter

Most problems in image restoration are ill-posed, so regulation technique is needed to restrict the problem. In this paper the err cost function with adaptive selection of regularization parameter (ASPR) is constructed in spatial domain, and the conjugate gradient is introduced to minimize the err cost function. In the frequency domain two constraints are incorporated in the estimation process of the object image and PSF. The proposed ASPR method can obtain the regularization parameter adaptively according to the edge information of the image which guarantees the restored image is the best result in the total field. Simulation results show that this method is correct and feasible, as well as has a good performance in the uniqueness and convergence of solution.