心电基线漂移去除方法的比较研究

A Comparative Study on Removal Methods of ECG Baseline Wandering

基线漂移可严重影响心电信号的分析与处理。滤波法和拟合基漂法是去除心电基线漂移的两类主要算法。本文讨论了形态滤波、中值滤波、线性相位FIR滤波、小波变换、形态和小波相结合、中值滤波和小波相结合等去除基漂的滤波方法,并与导数法与坐标法这两类基漂拟合方法进行了对比。实验结果表明,滤波法实现简单,但精度较低;分段拟合法去除基漂的效果更好,有自适应的特性而且失真较小。无论是实时性还是准确性,基漂拟合法都比滤波法更有优势;但某些场合下,基漂拟合点的提取十分困难,此时只能使用滤波法。本文比较了基漂去除的常用方法,从而为基漂去除方法的选择提供参考。

Baseline shift has severe effects on the analysis and processing of ECG. Filter and baseline polynomial-fitting are the main methods for removing baseline interference. This paper discussed the filtering ways of morphology filter,median filter,Linear-phase finite impulse response （FIR） filter,wavelet transform,morphology and wavelet combined method,median filter and wavelet combined method,comparing with the approach of baseline polynomial fitted way which composed of differential method and coordinates method. Experimental results showed that the poly-fitting approach are more efficient, had certain adaptive quality and little distortion, while filter ways were easy to implement but not accurate enough. As for real time and accuracy, polynomial fitted way has an advantage over filtering ways. Under the circumstances of having great difficulty in extracting baseline points, it is more appropriate to use filter ways to eliminate baseline wandering. This paper compared the methods in common use for suppressing baseline wander so as to help users to choose the proper way.