Nonlocal orientation diffusion partial differential equation model for optics image denoising

Abstract

In order to increase the ability of image denoising and protect the detail of image edge, a nonlocal orientation diffusion partial differential equation model for optics image denoising is proposed in this paper. Firstly, on the basis of P–M diffusion, we use nonlocal mean of gradient to control diffusion velocity and obtain the diffusion model of nonlocal gradient. Secondly, we let diffusion course go along with only such direction that can parallel to the image edge. What is more, we can restrain the diffusion of gradient direction. Finally, experiment contrast between our model and other local partial differential equation models is executed, the results of experiment indicate that our model can reduce the image noise better and preserve more edge features.