A new approach to robust blind restoration of motion blurred images with high fidelity

Abstract

In view of features of motion blurred images, a new robust blind restoration method with high fidelity is proposed in this paper, which can preserve image details and textures effectively. Compared with other restoration methods for motion blurred images, the new approach is characterized with a good property that no prior experience knowledge required. Meanwhile a new robust cepstrum-based PSF identification method is combined with a fast robust total variation-based restoration method to come into being an effective algorithm provided with high fidelity. A series of experiment results with artificial blurred images and natural blurred images verify the outstanding processing effect.