Research On Image Denoising And Amplification Algorithm Based On Adaptive And Fractional Order

With the rapid development of modern communication technology and mobile Internet,digital information processing capacity grows with the increase,the characteristics of digital image with its high transmission speed and large amount of information,has received wide attention in the field of image processing,the variational technique has also been fully used.Many very good results have been achieved,and many classic models and algorithms have been produced.In this paper,the algorithms and models of the related image processing are discussed by means of fractional order,adaptive operator and total generalized two order variational method.The main work is as follows:1.for additive Gauss removal of the traditional total variation model is prone to "staircase effect" and the fuzzy edge,this paper proposed a new total variation to the additive noise model,the new model is introduced and the Adaptive Fractional operator,numerical solution for the new model,the numerical algorithm is novel and the numerical results show that the new model not only has a good effect in inhibiting the "step effect",and at the edge of the protection is also very good.2.for the difficult removal of multiplicative noise,the total variation elimination multiplicative model is difficult to solve,and it is prone to "staircase effect " and edge blur.In this paper,a new variational adaptive de Poisson noise model combined with fractional order is presented.Based on the characteristics of Poisson probability distribution,the new model deduces the non convex adaptive regularization term.The new regularization can change the coefficients in the model adaptively according to the different regions of the image,so as to keep the edge of the image.In addition,this paper combined the fractional image information can be combined with more a bit,will be constant in the original model of first order differential with fractional differential substitution,it can achieve the purpose of removing the "staircase effect",the numerical solution of this model,this paper gives the primal dual and iterative numerical algorithm of the new.Finally,the experimental results show that the new model has obvious advantages compared with the traditional model,while protecting the details of the image,the ladder effect is also well suppressed.Besides,the new numerical method not only has the advantages of simplicity and fast convergence,but also performs better than the traditional numerical methods such as partial integral equation and Chambolle projection.3.In the same image enlargement model traditional total variation is derived based on the above shortcomings still exist based on considering non convex two order total generalized variation in the removal of the "step effect" and the fractional order in the protection of the detail information of image features,non convex total two order generalized variational model of Fractional Image magnification the order is put forward based on texture decomposition.The new model decomposes the low resolution image into two parts: cartoon and texture.Because the cartoon part represents the flat area of the original image,the traditional total variation is easy to produce "staircase effect".Therefore,the non convex two order general generalized variational model is applied to the cartoon part.The fractional order variational model is used to improve the detail information of the image,aiming at the texture part with more details of the image.For the new fractional order non convex new model,this paper uses the original dual and Chambolle projection algorithm to solve the numerical solution.The numerical results show that the proposed model has a significant improvement in the peak signal to noise ratio and mean square error.