摘要
抑制信号中的噪声干扰,是心电(ECG)信号预处理中的关键步骤。针对传统形态学滤波损失有用信号的缺陷,本文提出了一种基于形态学运算和自适应阈值的ECG信号消噪算法。首先,对含噪ECG信号进行形态学滤波和形态学峰谷提取运算;然后,估算形态学峰谷信号中时变噪声的即时方差,并依据3σ准则对峰谷信号进行自适应阈值处理,保留其中的有用信号;最后,将阈值处理结果与形态学滤波结果相加,作为ECG信号消噪处理的最终结果。仿真试验与实际应用结果表明,该算法不仅可以有效去除时变噪声的干扰,而且较好地保持了ECG信号的特征形态,处理效果明显优于以往的形态学滤波算法,且比基于平稳小波变换的消噪算法更适用于非平稳ECG信号的消噪处理。
Noise suppression is one of the important steps in ECG signal preprocessing. A new method based on morphological op- erations and adaptive threshold is presented to reduce signal loss by morphological filter for ECG signal de-noising in this paper. First, morphological filtering operation and peak-valley extracting operation are carried out in noisy ECG signal. Then, with the result of mor- phological peak-valley extracting operation, time-varying noise variance is estimated, and adaptive threshold processing based on 3or rule is applied to the peak-valley signal to preserve characteristic signal. Finally, the de-noising result of ECG signal is obtained by combining the reserved peak-valley signal with the result of morphological filtering operation. The results of simulation experiment and practical ap- plication show that the proposed method can not only suppress time-varying noise efficiently but also preserve the primary characters of ECG signal preferably. It performs better than the morphological filter method obviously. Moreover, it is more suitable for nonstationary ECG signal de-noising than the method based on stationary wavelet transform.
出处
《信号处理》
CSCD
北大核心
2009年第1期6-10,共5页
Journal of Signal Processing
关键词
心电信号
消噪
形态学运算
自适应阈值
ECG Signal
De-Noising
Morphological Operation
Adaptive Threshold
