[1]张业宏,陈恩平,么跃轩,等.基于双边滤波与离散余弦变换的NLM去噪算法[J].燕山大学学报,2018,42(3):259-264.[doi:10.3969/j.issn.1007-791X.2018.03.011]
 ZHANG Yehong,CHEN Enping,YAO Yuexuan,et al.NLM denoising algorithm based on bilateral filtering and discrete cosine transform[J].Journal of YanShan University,2018,42(3):259-264.[doi:10.3969/j.issn.1007-791X.2018.03.011]
点击复制

基于双边滤波与离散余弦变换的NLM去噪算法
分享到:

《燕山大学学报》[ISSN:1007-791X/CN:13-1219/N]

卷:
42
期数:
2018年第3期
页码:
259-264
栏目:
信息与计算机技术
出版日期:
2018-05-31

文章信息/Info

Title:
NLM denoising algorithm based on bilateral filtering and discrete cosine transform
文章编号:
1007-791X(2018)03-0259-06
作者:
 张业宏1陈恩平2*么跃轩12刘宝华2
1. 燕山大学 河北省并联机器人与机电系统实验室,河北 秦皇岛 066004;2. 燕山大学 机械工程学院,河北 秦皇岛 066004
Author(s):
 ZHANG Yehong1CHEN Enping2YAO Yuexuan12LIU Baohua2
 1.Hebei Provincial Key Lab of Parallel Robot and Mechatronic System,Yanshan University,Qinhuangdao,Hebei 066004,China; 2.School of Mechanical Engineering,Yanshan University,Qinhuangdao,Hebei 066004,China
关键词:
图像去噪双边滤波空间邻近函数离散余弦变换非局部均值
Keywords:
image denoisingbilateral filteringspace adjacent functiondiscrete cosine transformnon-local mean
分类号:
TP391.41
DOI:
10.3969/j.issn.1007-791X.2018.03.011
文献标志码:
A
摘要:
针对非局部均值去噪算法(NLM)易造成图像边缘模糊问题,提出了一种基于双边滤波和离散余弦变换的改进算法。该算法将双边滤波中的像素空间邻近函数与NLM算法的权值函数相结合,提出新的权值计算公式进而保护图像细节;利用离散余弦变换能量集中特性来计算像素相似性权值进而提高运算速度。首先将图像分割成子块,对子块进行离散余弦变换,然后在得到的离散余弦变换系数矩阵中筛选数据,最后用新权值计算公式在经筛选的离散余弦变换系数矩阵中度量像素的相似性。实验结果表明,与原NLM相比,该算法更好地保护了图像边缘细节特征和结构信息,峰值信噪比最大提高了1.4 dB,证明本文的算法去噪效果更佳。
Abstract:
Aiming at the problem of calculating slowly and bringing about edge fog in Non-Local Means (NLM) algorithm for image denoising, a kind of improved algorithm combined bilateral filtering (BF) with discrete cosine transform (DCT) denoising method was proposed in this paper. Space adjacent function of BF is coupled with original NLM weight function to protect image detail information. And focused power character of DCT is used to calculate the pixel similarity for improving calculation speed. Image is divided into several patches in this algorithm firstly and each patch is transformed by DCT. Then filtering data in the coefficient matrix of DCT. New weight function is used to calculate the similarity in coefficient matrix of DCT. Experiment results show that compared with NLM algorithm, the proposed method improves the calculation rate and protects image edge and detail information. The PSNR can be improved by 1.4 dB at most

备注/Memo

备注/Memo:
 收稿日期:2017-09-22    责任编辑:孙峰
基金项目:河北省自然科学基金资助项目(E2017203156);河北省高等学校科学技术青年基金资助项目(QN2017152);河北省博士后项目(B2016003021)
作者简介:张业宏(1990-),男,河北沙河人,硕士研究生,主要研究方向为图像处理与识别;*通信作者:陈恩平(1964-),男,黑龙江佳木斯人,博士,副教授,主要研究方向为自动化技术装备,Email:chenp@ysu.edu.cn
更新日期/Last Update: 2018-07-16