The Study Of A Number Of Issues On The Image Denoising Based On Partial Differential Equations

From the mid-1990 s, image noise reduction method of based on partial differential equations(Partial Differential Equations, abbreviated as, PDE) have started become a research hotspot in the field of image processing.Obvious de-noising effect and wide scope of application attracted the attention of many researchers at home and abroad. Under growing image data, the method effectively de-noising, while better able to retain the edge of the image information such as image segmentation, image enhancement, image restoration have achieved good results.This article focuses on the image de-noising issues that based on partial differential equations, the main contents of the paper are summarized as follows:Firstly,This paper introduced the purpose and significance of use of partial differential equations in image denoising field, research situations, analyzed some denoising models and advantages and disadvantages of the model’s coefficient of based on the partial differential equations. First studied the role of diffusion coefficient on the image denoising, in the process of image denoising by partial differential equation P-M model, the choice of diffusion coefficient will influence on the image de-noising effect, for which we proposed a new model of diffusion coefficient for image denoising. We discussed the importance of selecting the diffusion coefficient and gradient threshold on P-M model for image denoising, and compared the advantages and disadvantages of the two diffusion coefficients: according to their basis we proposed a new diffusion coefficient and applied to the CLMC discrete numerical model experiments. Image denoising can be effectively carried out by the proposed diffusion coefficient. Secondly, to study the advantages and disadvantages of anisotropy Total Variation(TV) diffusion equation model and isotropic thermal diffusion(TD) equation model to image denoising, in order to supplement their shortcomings we have proposed an improved adaptive hybrid model proposed which according to the image information can be adaptive for each area. More flat region of the diffusion model image, and less diffusion at the edge of the image, get a better denoising effect, while retaining edge. We compared the effects of three models for image denoising.Finally, we use two methods within MATLAB environment simulation tests were conducted. The test results show that SNR and PSNR that obtained by new diffusion coefficient better than SNR and PSNR that obtained by the diffusion coefficient from two P-M equations, therefore, not only the coefficient proposed It has characterized reservations at the edge of the image, and can have a good denoising effect.Improved adaptive hybrid model for image denoising can be efficiently performed.