结合高阶微分的非局部均值图像去噪算法  

Non-Local Means Algorithm Combined with High Order Differential

作　　者：彭扬[1] 

机构地区：[1]重庆大学数学与统计学院,重庆401331

出　　处：《计算机系统应用》2014年第1期119-123,共5页Computer Systems & Applications

摘　　要：非局部均值算法是近年来去噪效果优秀并且引起广泛研究的算法,但是该算法对于弱纹理,弱边缘不能很好地保持.在非局部均值算法的基础上,利用高阶微分可以很好地反映图像的弱边界和振荡等特征的特点,改进了非局部均值算法.改进的算法能够在去噪的同时保留更多的细节、纹理和弱边缘.Non-local Means algorithm is a very effective and widely studied algorithm, but it cannot maintain the texture and weak edges of image very well. By using the characteristics of high order differential which can reflect the weak boundary and oscillation characteristics of image well, we improve the Non-local Means algorithm. The improved algorithm cannot only remove noise, but also keep more details, textures and weak edge.

关 键 词：图像去噪 非局部均值 HESSIAN矩阵 高阶微分 特征方向 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]