A mixed single image motion deblur method

Abstract

The blind recovery of images is a difficult problem in the field of digital image processing. Researching the restoration of motion blurred images，a hybrid single image motion deblur method is proposed. Firstly, the PSF is estimated, then the latent image is evaluated by PSF and original blurred image, and the edge restoration, fuzzy kernel estimation and deconvolution are iterated several times. The adaptive unblind deconvolution method is used to obtain high quality deblurred image. Finally, compared with the experimental results of two classical algorithms, Wiener Filter and Lucy Richardson, the experimental results show that the proposed algorithm has strong ability to maintain details and better deblurring effect.

Keywords

• motion blur
• image restoration
• edge restoration
• fuzzy kernel estimation