变分模态分解在去除心电图信号基线漂移中的应用

Application of variational mode decomposition in removing ECG signal baseline drift

心电图(ECG)信号可反映心脏电生理功能活动状态并能为临床心脏病准确诊断提供有重要价值的信息。检测到人体的心电图信号,通常伴随着很多干扰,其中基线漂移干扰很大程度影响了对心电图信号的准确判断。根据变分模态分解(VMD)理论,提出了一种去除心电图信号基线漂移的方法。选择合适的分解层数,利用变分模态分解将心电图信号分解为一组模态分量,去除含有基线漂移成分的模态分量,重构剩余模态分量得到去除基线漂移后的心电图信号。通过ECG信号仿真和实际数据处理实验,与经验模态分解算法比较信噪比提高了4 d B,且能够保持心电图信号的形态特征,有效去除了基线漂移干扰。

ECG signal can reflect the state of cardiac electrophysiological function and provide important information for the accurate diagnosis of clinical heart disease. The ECG signal of human is detected usually accompanied by a lot of interferences. The baseline drift interference seriously affects the accurate judgment of ECG signal. Based on variational mode decomposition( VMD) theory,a method removing ECG signal baseline drift was proposed. By choosing appropriate parameters,the ECG signal was decomposed into a set of modal components by the variational mode decomposition. The modal component of the baseline drift was removed. The residual modal components were refactored,and the ECG signal was got after removing the baseline drift. Through the simulation of ECG signal and experiment of actual data processing,compared with the empirical mode decomposition algorithm,the signal noise ratio wass improved by4 d B. The morphological characteristics of the electrocardiogram was maintained. The baseline drift signal was effectively removed.