PSF-constraints based iterative blind deconvolution method for image deblurring

Abstract

In recent years, Image Deblurring techniques have played an essential role in the field of Image Processing. In image deblurring, there are several kinds of blurred image such as motion blur, defocused blur and gaussian blur. Many methods to address this problem have been proposed by researchers in previous research, among which the iterative blind deconvolution (IBD) method is the most popular method to solve this problem. However, the convergence of this method is not ensured, and there is no effective method to choose a proper initial estimate image and PSF(point spread function). In this paper, we improve the iterative blind deconvolution method by adding several constraints, which could be the type or parameters range of the PSF, on the PSF in each iteration. Experiment results validate that, with the help of these newly added constraints, our method are more likely to converge and has better deblurring performance than the IBD. © 2010 Springer-Verlag Berlin Heidelberg.