一种改进的各向异性扩散模型

A New Modified Anisotropic Diffusion Equations

各项异性扩散方程是一种经典的图像去噪方法,但该方法在去除噪声的过程中会造成一定程度的模糊边缘。对此文中提出了一种基于改进的各向异性扩散方程的图像去噪方法,通过在其能量泛函的目标函数中添加残差项,使能量泛函的极小解更加接近原始的函数,可取得比其更好的去噪效果。文中方法可看作是各项异性扩散方程和全变差模型的结合。实验表明,新提出的方程相对经典的方程有较好的边界处理效果和更高的信噪比。

The anisotropic diffusion equation is a classical image denoising method, but this method can cause a certain degree of fuzzy edge in the process of removing noise. This paper presents a based on improved anisotropic diffusion equation for image to denoising method, through in the objective function of the energy functional add residuals and make the energy functional minimizer is closer to the original function can get better denoising effect. The method can ments show be regarded as the combination of the anisotropic diffusion equation and the total variation model. Experithat the new equation has better boundary treatment effect and higher signal to noise ratio than the classical equation.