带反馈修正的多结构元形态学心电信号QRS波检测算法

A Multi-Structuring-Element Morphological Approach with Feedback Correction for Detecting the QRS Complex in ECG Signals

本文提出了一种带反馈修正的多结构元形态学心电信号QRS波检测算法。首先,采用形态学滤波消除心电信号中的脉冲噪声;接着,基于两种不同宽度的结构元,对滤波后信号进行形态学峰谷提取,获取QRS波波形陡峭信息和幅度信息;然后,根据形态学峰谷提取结果,采用自适应阈值检测法初步定位QRS波;最后,利用估算出的R-R间期对QRS波检测结果进行评估和修正,并反复此评估修正过程直至获得满意结果。经MIT-BIH心律失常数据库验证,本算法具有很好的QRS波检测效果。

This paper describes a multi-structuring-element morphological approach with feedback correction for detecting the QRS complex in ECG signals. First, impulsive noise is removed from the input ECG signals by the morphological filtering operation. Second, the morphological peak-valley extraction operation is employed to the filtered signals with two structuring elements of different sizes to obtain the information of the QRS complex about the steep shape and amplitude. Then,the QRS complex is located using the adaptive threshold detection. Finally, the detection results are evaluated and corrected according to the calculated RR interval, and the evaluation-correction process is repeated until the obtained detection results are satisfied. The proposed approach has been tested using the MIT-BIH Arrhythmia Database. And the results show that it performs well in detecting the QRS complex.