基于EEMD的脉搏信号改进阈值去噪研究  

ON EEMD-BASED IMPROVED THRESHOLD DENOISING FOR PULSE SIGNAL

作　　者：刘攀[1] 夏春明[1] 燕海霞[2] 王忆勤[2] 郝一鸣[2] 徐琎[2] 许文杰[2] 

机构地区：[1]华东理工大学机械与动力工程学院,上海200237 [2]上海中医药大学基础医学院,上海201203

出　　处：《计算机应用与软件》2016年第3期76-79,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(81173199;81102729)

摘　　要：针对脉搏信号的非线性、非平稳特性,及其干扰源的分布特点,提出一种基于聚合经验模态分解(EEMD)和小波阈值去噪的改进算法。根据脉搏信号在各固有模态函数(IMF)上的分布特点,在有效滤除高频干扰的同时,采用网格搜索对低频IMF分量进行阈值选取去噪,有效去除其低频噪声,实现自适应且有效的脉搏信号去噪处理。仿真与实测结果表明,基于EEMD的改进阈值去噪算法可有效滤除脉搏信号中常见的白噪声、工频干扰、基线漂移、呼吸效应,明显改善了脉搏信号的去噪效果,且极大程度地保留了脉搏信号的内在性质,为脉搏信号预处理提供了一种有效手段。Considering the non-linear and non-stationary characteristics of pulse signal and the distribution feature of its interference sources,we proposed an improved denoising algorithm which is based on ensemble empirical mode decomposition（ EEMD） and wavelet threshold. According to the distribution feature of pulse signal in each intrinsic mode function（ IMF）,it carries out the threshold selection denoising on IMF component of low frequency with grid search while effectively filtering the high frequency interference,which effectively removes its low frequency noise,and realises the adaptive and effective pulse signal denoising processing. Results of simulation and actual measurement showed that the EEMD-based improved threshold denoising algorithm could effectively filter out 4 kinds of interferences common in pulse signals： the white noise,the power-line interference,the baseline drift,and the respiratory interference,it obviously meliorated the denoising effect on pulse signal and retained the intrinsic nature of pulse signal to a great extent,this provided an effective means for preprocessing the pulse signal.

关 键 词：脉搏信号 EEMD 阈值 去噪 网格搜索 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]