心电信号波形保持的分数阶零相位滤波

Electrocardiogram Signal Waveform Preserved Fractional Zero-Phase Filtering

针对经典滤波方法难以兼顾保真与去噪,易产生相位失真等问题,设计了一种基于分数阶Liouville和Weyl积分算子的并联型分数阶零相位滤波器.基于分数阶微积分的幅频特性与相频特性均受控于分数阶次的特点以及反向滤波的反向相频特性,将零相移滤波思想植入分数阶滤波器设计,提出适用于心电信号去噪的分数阶零相位滤波算法.基于MIT-BIH心律失常数据库的心电信号,分别从定性和定量两方面进行MATLAB仿真实验.仿真结果表明,分数阶零相位滤波器在滤除心电信号噪声的同时,能较好地保留原始信号的主要特征波形,有效避免相位失真,具有保真性能好、抗噪能力强、计算复杂度低的优点.验证了所提算法的有效性.

The idea of zero-phase filtering into the design of fractional filter was introduced and a novel fractional zero-phase filtering algorithm was proposed which better applies to Electrocardiogram（ECG） signal de-noising. The proposed filtering algorithm has the ability to better preserve the main features of ECG waveforms while filtering out noises, and effectively avoids the occurrence of phase distortions, demonstrating the advantages of better fidelity performance, strong noise immunity and lower computational complexity. Simulation experiments were conducted in MATLAB environment using testing ECG signals from the MIT-BIH Database. The results of the study demonstrated superior performances on visual sense and quantitative comparisons.