Research On Image Inpainting Algorithm Based On Fractional Order Differential

Image restoration is one of the important research issues in the field of computer vision and image processing.It has important and extensive applications in the field of cultural relics restoration and protection.For example,precious photographs and paintings are damaged to a certain extent due to various factors such as storage time,preservation methods,human destruction and so on.Image restoration technology refers to filling the damaged area according to certain rules according to the known information around,so that the restored image approaches or achieves the visual effect of the original image.Therefore,image restoration has important research significance in computational vision and image processing.TV(Total Variation Model)repair model not only has good repair effect,but also has excellent edge-preserving performance,so it has been widely used in image restoration.However,the regular term of TV model is the first derivative,which easily leads to the blurring of texture features with weak derivative properties,thus affecting the accuracy of repair.This paper mainly focuses on the restoration of weak edges and texture images.The main work is as follows:1.The traditional TV model is prone to blur and ladder effect in repairing damaged images.To solve this problem,a new TV model restoration algorithm based on improved diffusion coefficient is proposed.A new diffusion coefficient is added to the original TV model.This coefficient can not only ensure that the diffusion coefficient of the image at the large gradient modulus is greater than zero,but also solve the problem of TV model diffusion stopping at the large gradient and avoid the restoration falling into local optimum.In the process of image restoration,edge region diffusion is small and smooth region diffusion is large,which overcomes edge blurring and ladder effect in image restoration,so as to improve the accuracy of image restoration.Experiments on rock images with weak textures and edges show that the accuracy of the improved model is improved.2.Aiming at the problem that the preservation effect of TV model on texture details and other information is not ideal,and that it is difficult to calculate regularization and data items in the process of model minimization because they are not differentiable at the origin,an image restoration algorithm combining texturestructure information and fractional TV model is proposed.When solving the fractional TV model,a small parameter is added to the gradient calculation,which overcomes the problem that the regular term and data item can not be differentiated at the origin,and increases the stability of the model,so that the weak edge and texture of the image can be well preserved in the process of image restoration.In addition,the improved model is based on the graph.As the prior information of the known region determines the texture direction of the region to be repaired,the texture details and weak edge information in the image are better preserved.Both theoretical analysis and experimental results show that the improved model can solve the problem that the TV model is not ideal for weak edge and texture image restoration,and can determine the direction of image restoration based on the prior information of the known region of the image,and the efficiency of image restoration has been improved.