一种非线性复扩散与冲击滤波的图像消噪方法

摘要

In the complex diffusion, the imaginary part of image serves as approximate second derivative of image, uses the imaginary part to control the diffusion and can get the forward and backward complex diffusion. But the forward and backward complex diffusion can' t denoise or maybe enlarge the noise of detail areas. A novel forward and backward complex diffusion process combined with shock filter is proposed in this paper. In the process, the complex diffusion is utilized to denoise the whole image and shock filter to deblur the detail. In the experiments, the results demonstrate that the novel dif fusion can denoise the image, retain the details effective and get higher signal to noise ratio.%由于图像复扩散的虚数部分等价于图像高斯卷积后二阶梯度算子,因此利用复扩散的虚部去控制扩散系数可得到前后向复扩散,然而这种选择性的前后向复扩散不能有效消除图像细节区域的噪声.针对前后向复扩散的缺点,提出一种复扩散与冲击滤波相结合的前后向复扩散,利用复扩散去消噪,虚部控制冲击滤波器实现图像边界的锐化.通过实验证明了这种扩散消噪的同时更好的保留图像的细节,且具有更高的峰值信噪比.