基于平稳小波变换心音信号阈值去噪的对比分析  

Comparison and Analysis of Heart Sound Signal Threshold Denoising Based on Stationary Wavelet Transform

作　　者：高清河[1,2] 刚晶[1] 王和禹[1] 

机构地区：[1]辽宁中医药大学信息工程学院,沈阳110847 [2]东北大学理学院,沈阳110004

出　　处：《计算机与数字工程》2015年第6期957-959,共3页Computer & Digital Engineering

基　　金：辽宁中医药大学非医药专业项目(编号:601152)资助

摘　　要：为研究去除心音信号中的环境噪声、工频噪声和肌电噪声等多种噪声,获得最优的心音去噪方法。论文基于平稳小波变换选取db5小波对心音信号进行3尺度分解,对分解后的高频系数分别选择了硬阈值函数以及软阈值函数,并对固定阈值、Stein无偏似然估计阈值、启发式阈值、极大极小阈值以及Penalized阈值进行了对比分析。通过去噪衡量指标对心音信号不同阈值以及阈值函数的去噪效果对比分析,硬阈值的去噪效果要好于软阈值的去噪效果,去噪效果最好的是Stein无偏似然估计阈值,有效地去除了参杂在心音信号中的多种噪声。To remove three kinds of noise in heart sound signal including environmental noise、powerline interference and electromyographical interference and obtain the optimal denoising method of heart sounds,based on stationary wavelet transform,heart sound signal is decomposed into three layers using db5 wavelet.Hard threshold function and soft threshold function are applied to high frequency coefficients respectively.Then,the comparison and analysis of denoising effect are carried by Sqtwolog,Rigrsure,Heursure,Minimaxi,Penalizd high,Penalizd medium and Penalizd low threshold.Hard threshold denoising effect is better than the soft threshold and the Rigrsur threshold method is the best in all the thresholds.Rigrsur threshold method based on stationary wavelet transform can achieve good denoising effect and effectively remove variety of noise of heart sound signal.

关 键 词：心音信号 平稳小波变换 阈值函数 阈值 阈值去噪 

分 类 号：R318[医药卫生—生物医学工程] TP391[医药卫生—基础医学]