中文摘要：

针对心电信号中混有的基线漂移、工频干扰、肌电干扰等噪声,比较了适于心电信号的4种基于小波变换的心电信号消噪算法,结合消噪后的信噪比和信号失真度,提出一种综合小波变换的心电信号消噪算法。该算法先使用小波分解法消除心电信号中的基线漂移,再利用模极大值法消除工频干扰、肌电干扰等噪声。并且运用该算法对MIT-BIH心律失常数据库中的含有多种噪声的心电数据进行了仿真与实验,结果表明噪声被有效地消除并且失真度较小,可满足临床分析与诊断对心电波形的要求。

英文摘要：

The quality of electrocardiogram （ECG） is often compromised by baseline wandering, power line interference and muscle noises. This paper compares four wavelet transform （WT） algorithms and proposes a combined algorithm to remove noises from ECG signals. The combined algorithm generates higher signal to noise ratio （SNR） and lower distortion factor than the four individual algorithms. The combined algorithm adopts the WT decomposition at high-scale to restrict BW and wavelet transform modulus maximum to reduce the rest noises. We tested the proposed algorithm through MIT-BIH arrhythmia database. The final results confirm that the proposed algorithm effectively eliminates noises and gives little distortion, and the filtered ECG waveforms can be used for clinic analysis and diagnosis.