Based On The Total Variation Model Image Restoration Research And Applications

Digital image inpainting plays an important role in image processing and has a broad application fields. It is also the research focus in the fields of image processing and computer vision . The thesis focuses on research and application of image inpainting based on PDE.The classical BSCB, CDD and TV model are studied firstly. The characteristics of various models are summarized by analyzing the theory and comparing the experiment results. Baaed on the TV inpainting model ,the cause of the problem that the connectivity is destroyed is analyzed and an improved scheme is also proposed .The main idea of the scheme is the introduction of the distance function in the TV inpainting model and the TV inpainting weighting by using the distance function. We can treat the contribution of the total variation differently according to the position and the different orientation of the gradient ,thus realizing the inpainting. The introduction of the weighted factor not only can solve the problem of the connectivity but also can better remain the advantages of denoisng and the inpainting efficiency of the original TV model. The inpainting experiments are completed after the theory analysis and the research of the numerical realization .The results are satisfactory and can achieve the aim of the improvements.The improved weighted TV inpainting model is extended to the inpainting of the vector image. The method draws lessons from the idea of the BC_TV and completes the inpainting of the vector image by the combination of the weighted TV model and the BC_TV model. The experiments prove that the inpainting results are satisfactory.Based on the above research, we can make sure that beneficial exploration is carried out in the digital image inpainting based on the PDE and some valuable research achievements are made.