基于EMD与IMF分量统计特性的ECG去噪  

ECG signal denoising based on EMD and statistical characteristics of IMF components

作　　者：卢莉蓉[1] 牛晓东 王鉴[3] 李春燕 LU Lirong;NIU Xiaodong;WANG Jian;LI Chunyan(Department of Biomedical Engineering,Changzhi Medical College,Changzhi 046000,China;Department of Basic Medicine,Changzhi Medical College,Changzhi 046000,China;Key Laboratory of Information Detection and Processing,North University of China,Taiyuan 030051,China;Process Technology Research Institute,Shanxi North Machine-Building Co.,Ltd,Taiyuan 030051,China)

机构地区：[1]长治医学院生物医学工程系,山西长治046000 [2]长治医学院基础医学部,山西长治046000 [3]中北大学信息探测与处理山西省重点实验室,山西太原030051 [4]山西北方机械制造有限责任公司工艺技术研究所,山西太原030051

出　　处：《中国医学物理学杂志》2021年第12期1529-1534,共6页Chinese Journal of Medical Physics

基　　金：国家自然科学基金(61842103);山西省高等学校科技创新项目(2020L0389)。

摘　　要：经验模式分解(EMD)域内心电(ECG)信号的去噪,通常为基于QRS特征波经验性识别固有模态函数(IMF)分量并重建ECG信号。由于该方法引入个人误差,因此识别不准确。针对此问题,本文提出利用EMD与IMF分量统计特性对ECG信号进行去噪。本方法首先对含噪ECG信号进行EMD分解得到一系列IMF分量,然后利用IMF分量的统计特性识别IMF分量属性,并采用被识别为ECG信号的IMF分量重建ECG信号。该识别方法基于统计学方法,具有统计学和现实物理意义。将本方法应用于真实ECG信号去噪处理中,结果表明,本方法可有效去除ECG信号基线漂移噪声与肌电干扰噪声,去噪效果优于经验法。The denoising of electrocardiogram(ECG)signal in empirical mode decomposition(EMD)domain usually involves the empirical identification of the intrinsic mode function(IMF)component based on QRS characteristic wave and the reconstruction of ECG signal.However,due to the personal errors caused by empirical method,the identification is inaccurate.To solve this problem,EMD and the statistical characteristics of IMF component are used to denoise ECG signal.Herein a series of IMF components are obtained by EMD on noisy ECG signals,and then the properties of IMF components are identified using the statistical characteristics of IMF components,and the IMF components identified as ECG signals are used to reconstruct ECG signals.The proposed identification method is based on statistical method,with statistical and practical significance.The proposed method is applied to real ECG signal denoising,and the results show that the method can effectively remove the baseline drift noise and electromyography interference noise of ECG signal,achieving a denoising effect better than that of empirical method.

关 键 词：心电信号 去噪 经验模式分解 固有模态函数分量 

分 类 号：R318[医药卫生—生物医学工程] TN911.7[医药卫生—基础医学]