基于保边函数的图像去噪

Image de-noising based on edge-preserving function

为解决传统去噪算法对图像平滑区域效果较好,但存在边缘模糊且残余噪声较大的不足,在分析传统各向同性扩散的基础上,以去噪目的为先验知识建立保边去噪模型,提出了基于保边函数的图像去噪算法.该算法对图像平滑区域进行各向同性的平滑处理,保留了传统算法的优点;对图像边缘区域进行各向异性处理克服了传统算法的边缘模糊现象.实验结果表明:基于保边函数的去噪算法具有残余噪声较小和保边性,提高了图像的PSNR和视觉效果.但是对图像中的较小边缘误作为噪声被去除.

Traditional de-noising algorithms have a better effect in smooth area than that at the edge of image.There are still some shortcomings,such as edge blurry and large residual noise.This study has developed an edge de-noising model with de-noising purpose as prior knowledge based on the analysis of traditional isotropic diffusion.Also,this study has derived a new de-nosing algorithm based on edge-preserving function.The algorithm applies the isotropic smooth processing on smooth area to preserve the advantages of traditional algorithms.On the other hand,the algorithm uses the anisotropic processing on the edge of image to overcome edge blurry phenomenon.Experimental results demonstrate that the algorithm achieves a less residual noise and edge protection,improves image PSNR and visual effects.However,small edges are removed as noise.