Research On Image Denoising Based On Variation Method And Partial Differential Equations

In the past two decades, image processing based on Partial Differential Equations(PDEs) has made great progress and attracted the attention of many researchers. In this papar, we mainly research the image denoising models based on variation and partial differential equations and analyze these models. On the basis of previous knowledge, we put forward two improved models. Through the numerical experiments, we can see that the denoising effect of the improved models is better than that of the original models. Then two following questions are mainly solved, and some results are given as follows :Firstly, this paper introduces the improved weighted function of the NL-ROF denoising model. Through analysis of several weighted kernel function, this paper proposes a new cosine Gauss kernel function. The bilateral filter is a nonlinear filter which smoothes an image signal for denoising while preserving edges and details. Not only it takes into account spatial proximity also considers the similarity of intensity. So an improved weighted non-local means algorithm for image denoising is proposed. In this paper, the discrete form of the non-local means denoising algorithm is analyzed and improved. Through the numerical experiments, we can see that the denoising effect of the improved models is better than that of the original models.Secondly, this paper introduces the PM denoising model. The improved denoising model is proposed through adding a forced term. The anisotropic diffusion denoising model added common forced term can not protect the edges, so a new forced term is proposed through the analysis of the Alvarez denoising model. The proposed method has been evaluated on testing images with various levels noise. Numerical results show that the proposed method improves the denoising performance as well as the preservation of structure information.