The Research On Algorithm Of The Anisotropic Image Denoising Based On The Bending Edge And Complex Texture’s Keep

This paper mainly focuses a hot and difficult point in the field of image processing:in the process of image denoising, how to better preserve the bending edge and complex texture. On the one hand, in the process of image acquisition, access and transmission links, by the influence of various factors, susceptible to result lower of image quality, submergence of details and characteristics, deterioration of visual effects and so on.Finally, making the image have the lower value. At the same time, in the production and life of the today’s society, the applied requirements of the image will be more and more widely. On the other hand, in the process of image denoising, the blurry and smooth of the original image was inevitable. This is a contradictory with the preservation of the image’s details.Firstly, this paper provides an overview of the anisotropic diffusion model, analysis and summaries three types of classical discrete way about diffusion function in the existing of models, P-M anisotropic diffusion model P. Perona and J.Malik, consistency anisotropic diffusion model and forward-and-backward diffusion filtering model proposed by Weiekert’s. At the same time, theoretical research and experiments prove that these discrete methods could not be better preserved curved edges and complex texture, mainly is due to the diffusion could not spread along the edge bending.In order to overcome these defects, on the basis of the P-M anisotropic model, put forward by use of the time accumulation of stepping anisotropic diffusion model to denoising of the image. The diffusion model mainly adopts two mechanisms, these are the stepping move of the diffusion center and the time accumulation of diffusion points in the different directions. The first mechanism depends on the information of the image structure, to determine the direction of stepping move through the eight neighborhood difference operator. The diffusion points can better reflect the characteristics of initial diffusion center. The time accumulation of diffusion points in the different directions was achieved by our proposed the diffusion accumulation function. Appears to be a simple process of accumulation, while it realized the continuous optimization of the results, and get a stable and reliable value.In the end of this article, we give out the simulation experiments of the proposed method through MATLAB. The experiments proved that, our algorithm can better keep image edge bending and complex texture in denoising.