一种震动耦合扩散的去除乘性噪声模型

针对传统扩散模型在去噪中存在的边缘模糊问题,将震动滤波引入各向异性扩散模型中,建立了一种去除乘性噪声的模型。在去噪算法中采用BZN来区分边缘、噪声以及平滑区域。借助一种新的扩散速度函数,可以更好地去除乘性噪声。数值实验表明,该去噪模型可以更好地保护边缘,抑制＂阶梯效应＂。

A Shock Coupled Diffusion Model for Multiplicative Noise Removal

In view of the traditional diffusion model denoising problem to blur edge,the shock filter into the anisotropic diffusion model is introduced,and a model to remove multiplicative noise is established.The edge,noise and smooth areas are distinguished by use of the BZN in the denoising algorithm.By means of a new diffusion speed function,the multiplicative noise can be removed efficiently.Numerical experiments show that the proposed denoising model can preserve edges and suppress the＂staircase effect＂effectively.