基于边缘增强算子的各向异性滤波方法  

Anisotropic Diffusion Filtering Based on Edge Enhancement Operator

作　　者：朱丽琪[1] 范雪峰[1] 周志平[1] ZHU Li-qi,FAN Xue-feng,ZHOU Zhi-ping（Research Center of CAD,Tongji University,Shanghai 021804,China）

机构地区：[1]同济大学CAD研究中心,上海201804

出　　处：《电脑知识与技术》2011年第9期6214-6217,共4页Computer Knowledge and Technology

摘　　要：对CT图像进行有效的抑制噪声并保持边缘信息,有利于图像的后续处理。传统的各向异性扩散能有效地抑制噪声,但会丢失图像细节,边缘保持的效果也不理想。该文提出了基于边缘增强算子的各向异性滤波方法。首先,使用基于梯度直方图的各向异性扩散方法计算梯度门限;然后,在非线性各向异性扩散方程中增加一个边缘增强算子,来增强边缘信息,尽量少的丢失有用信息。实验结果表明,该方法在噪声消除和边缘保留方面能获得较好的效果。For the purpose of further processing like segment,it is necessary to remove noisy while keeping the edge＇s information in CT image processing.The conventional anisotropic diffusion is effective in removing noises,but it often destructs the structural information of the object in the image because of a sequence of inaccurate diffusions.A new anisotropic diffusion approach based on edge enhancement operator is proposed.At first,gradient threshold K is calculated by using GHAD（Gradient Histogram-Based Anisotropic Diffusion） for anisotropic diffusion.Secondly,an edge enhancement operator is incorporated into the diffusion steps,to obtain the edge details.Experiments on CT images indicate effective preservation of edge,removal of noise by our scheme.

关 键 词：医学图像滤波 各向异性扩散 边缘增强 

分 类 号：TP391[自动化与计算机技术—计算机应用技术;自动化与计算机技术—计算机科学与技术]