Research On Image Deblurring And Image Enhancement

Image deblurring and image enhancement are fundamental and widely studied subjects in image processing where the main purpose is to reconstruct the clean image from the degraded and/or noisy image.The main detrimental factors to corrupt the image are noise,data missing,and blur,to name a few.The main contributions of this thesis are to make full use of effective prior knowledge to restore the degraded images.The specific research contents are divided into three parts:1.The problem of image recovery under the case of salt-and-pepper noise and data missing is addresed.The salt-and-pepper noise reviewed as the impulsive noise is modeled as a sparse signal because of its impulsiveness.The data missing pattern is denoted by a sparse vector that contains only zeros and ones to formulate the data missing.To accurately reconstruct the clean image,the wavelet and framelet domains are exploited to sparsely represent the image.From the reformulations conducted and to recover the image,a joint estimation is devised to simultaneously perform the image recovery,the salt-and-pepper noise suppression and the missing patter estimation under a optimization framework.To solve the optimization problem,two efficient solvers based on the alternating direction method of multipliers(ADMM)and accelerated proximal gradient(APG)are developed to obtain the joint estimation solution.2.The problem of image recovery in the presence of salt-and-pepper noise and image blur is addresed.The salt-and-pepper noise is modeled as a sparse signal because of its impulsiveness.To accurately reconstruct the clean image and the blur kernel,the framelet domains are exploited to sparsely represent the image and the blur kernel.From the reformulations conducted,a joint estimation is devised to simultaneously perform the image recovery,the salt-and-pepper noise suppression and the blur kernel estimation under an optimization framework.3.The problem of blind image recovery using multiple blurred-noisy images of the same scene is addressed.To perform blind deconvolution,also called blind image recovery,the blur kernel and image are,respectively,represented by group sparse domains to exploit the local and nonlocal information so that a novel joint deblurring approach is thus conceived.In addition,the reweighted data fidelity is also developed to further improve the recovery performance,where the weight is determined by the estimation error.Moreover,in order to reduce the undesirable noise effect in group sparse representation,distance measures are studied in the block matching process for finding similar patches.In such a joint deblurring approach,a more sophisticated two-step interactive process is needed in that each step is solved by a means of the well-known split Bregman iteration(SBI)algorithm,which is particularly used to efficiently solve the proposed joint deblurring problem.