Digital Image Of The Blind Rehabilitation Research

With the rapid development of multimedia and computer technology, image processing becomes more and more widely used. Blind image restoration is a great part of image processing, which refers to the technology field that study how to restore the original image from the observed image without sufficient a prior knowledge about the degradation process. Blind image restoration has a strong background of mathematics, including estimation theory, ill-posed problem solution method, linear algebra, stochastic process, numerical analysis, and so on. The general sequence of blind image restoration is: firstly, model the degradation processing in mathematical form, and secondly, solve this inverse problem to get an estimation of original image.In this thesis, we do some research for the problem of blind restoration of gray image. Two methods of blind image restoration are discussed:Firstly, error-parameter estimation is discussed. This method is used to estimate point spread function(PSF) of blurred image which can be denoted by a parameter, such as linear motion model, gauss model. Then we use routine recovery algorithm to restore the blurred image, such as wiener algorithm.Secondly, NAS-RIF(nonnegativity and support constrains recursive inverse filtering) is discussed. The basic idea of NAS-RIF is given and by integrating the idea of regularization, an improved algorithm is given. Then effectiveness of this algorithm is proved by many simulations.