高斯平滑算子对Y-K模型改进的图像去噪方法  

Algorithm of Image Denoising Based on High Order Partial Differential Equation

作　　者：叶衍昌[1] 赵东红[2] 李小娟[1] 

机构地区：[1]华北计算技术研究所,北京100083 [2]北京科技大学数理学院,北京100083

出　　处：《软件导刊》2013年第11期161-164,共4页Software Guide

摘　　要：研究了在偏微分方程理论框架下进行图像去噪的方法,重点对二阶和四阶偏微分方程的主要去噪方法进行了分析。二阶偏微方程模型中的P-M模型能很好地去除噪声,但时常会出现块状现象;四阶偏微分方程模型中Y-K模型能消除阶梯效应但会出现斑点现象。使用Gilboa扩散系数、中值滤波器和高斯平滑算子对Y-K模型进行改进。实验证明,新模型减少了迭代次数,提高了算法效率,也一定程度上避免了块状及斑点现象。This paper studied the theoretical framework of the partial differential equations for image denoising method, the second and fourth order partial differential equations of the main denoising method were discussed and a new algo- rithm. The P-M model of the second order partial differential equations denoised well, but them led to a ＂massive phenom- enon ＂. The Y-K model of the fourth order partial differential equations eliminate the step effect but will appear the speck- le phenomenon. A new model introduced is proposed that the Y-K model is improved by using The diffusion coefficient of Gilboa, median filter and Gaussian smoothing operator. A large extent, the new model proved to reduce the number of it- erations, but also to some extent to avoid a ＂massive phenomenon ＂and the speckle.

关 键 词：图像去噪 偏微分方程 扩散系数 中值滤波 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]