基于B-样条双正交小波R波的标定和QRS波检测  

R waveform calibration and QRS waveform detection of B-splines-based biorthbogonal wavelets

作　　者：苏丽敏[1] 戴启军[1] 王杰[1] 

机构地区：[1]郑州大学电气工程学院,河南省郑州市450001

出　　处：《中国组织工程研究与临床康复》2009年第9期1657-1660,共4页Journal of Clinical Rehabilitative Tissue Engineering Research

摘　　要：实现心电信号QRS波检测的算法很多,文章给出了一种基于B-样条双正交小波对心电信号R波峰值标定和QRS波波段检测的方法。利用双正交样条小波等效滤波器,对心电信号按Mallat算法进行快速变换;从信号奇异点的李氏指数与模极大值关系的角度,分析心电信号奇异点(R峰值点)与其小波变换模极大值对的零交叉点的关系。用二次B-样条小波滤波器组对心电数字信号进行4个尺度的小波分解,然后根据分解的尺度波形特性求出正负极值对过零点,即R波峰值,并检测出QRS波段。采用Matlab编程实现该算法。从实验结果可以得出,该算法对心电信号中QRS波群的特征提取和几种常见的心电干扰具有较强的鲁棒性,经MIT-BIH标准心律失常数据库验证,QRS波的正确检测率达99.9%。文中给出了程序流程图。There are many ways to detect electrocardiogram （ECG） QRS waves. The paper introduced a method about R waveforms calibration and QRS waveforms eetection of Biorthbogonal wavelets based on the B-splines. Using Biorthogonal Wavelet Filter, the ECG signal was quickly transformed by Mallat algorithm. The signal singularity＇s Lipschitz exponent was used to analyze the relationship between the signal singularity （peak R） and the zero-crossing point of the modulus maximum pair of its wavelet transfer. Using B-spline wavelet filter decomposes ECG signal to four scale wavelet. According to decomposed wavelet scale, R wavelet peak value was obtained and QRS wave band was detected by Matlab method. Experimental results demonstrate that this method is robust to time varying characteristics of QRS complexity and noise. The correct detection rate of QRS complex is 99.9% over the MIT-BIH Arrhythmia Database. The flow chart of the program was also shown in this paper.

关 键 词：B-样条小波 双正交小波 R波峰值 QRS波 

分 类 号：R318[医药卫生—生物医学工程]