The Application Of Improved Anisotropic Diffusion Method In Image Denoising

The paper mainly studies the application of anisotropic diffusion in image denoising. In the process of image generation, transmission and communication, variety of noise often accompany with it, which will degrade the quality of image, how to effectively denoise becoming a research topic. In recent years, as the theory of partial differential equation became a new tool in image denoising, which cause the attention of people gradually. Its developing experiences the processes from linear to nonlinear, from the isotropic diffusion to the anisotropic diffusion, and expands the real number field to the complex domain.This article simply introduces several classic models based on partial differential equation in image denoising, and mainly studies Weickert's coherence enhancing diffusion model, in view of this model's advantages and disadvantages, the pretreatment and diffusion tensor are improved respectively, two different anisotropic diffusion denoising method are proposed, and the simulation be done with Matlab7.0. One anisotropic diffusion method is based on image feature, first the wiener filter is used to do image preprocessing, then the image coherence is utilized to judge the image feature, and the diffusion tensor's eigenvalues are set in the corresponding regional; Another method firstly use morphological close-open operator to do pretreatment, then structure tensor is designed by second directional derivatives, and eigenvalues of diffusion tensor are devised according to adaptive gradient threshold value.The experimental results show that the two improved methods all inherit the advantages of coherence enhancing diffusion model, they are capable of removing noise and preserving the details of image characteristics efficiently, eliminating the false edge in plain area commendably, achieving better effect in image denoising.