Image Denoising And Deblurring Problems Based On Total Variation Type Functional

Noise and blur often make us obtain degraded images, this will lead to some verydifcult image post processing wok (such as: image segmentation, feature extractionand target recognition, etc), so efective image denoising and deblurring have been theimportant research topics in the feld of digital image processing. The goal of imagedenoising and deblurring is to recover original images as well as possible, which makeborders clear and retain details from the degraded images. Therefore it has importanttheoretical signifcance and practical value to research the models and the correspondingmethods of image denoising and deblurring, which can efectively keep important structureinformation.In the frst chapter, we briefy introduce the research background and some prelimi-naries of image denoising and deblurring.In the second chapter, frstly, we propose a local adaptive total variation model whichnot only can smooth out the noise but also can keep edge, then we use the primal-dualmethod to solve it. At last, experiments show the availability of the model and the methodwe proposed.In the third chapter, for classic image deblurring problems, this paper puts forwardbilateral constraint variational regularized image to deblurring problems and gives alter-nate Bregman method to solve it. The experimental results show that for diferent levelsof gaussian blur and noise, the model and algorithm can get efective recovery results.