【摘要】 在心电信号的采集过程中，各种噪声的干扰会引起信号失真及基线漂移，进而影响对心脏信号的精准判断。针对此，提出一种基于集合经验模态分解的自适应算法。首先，对含有噪声及基线漂移的心电信号进行集合经验模态分解(EEMD),分解出固有模态函数(IMF)分量。然后，筛选出需要处理的IMF分量。最后，通过自适应窗口处理带噪的低阶IMF以及移除导致基线漂移的高阶IMF,从而达到降噪和修正基线漂移的目的。在MIT-BIH数据库中的实验结果表明，基于EEMD方法的降噪效果良好，在同等肌电噪声情况下，与基于EMD的自适应窗口法对比，在平均信噪比上提升1.750 7,增幅约为13%;在同等基线漂移情况下，与基于EEMD的阈值法对比，在平均基线矫正率上下降0.079 5,降幅约为14%。更多还原

【Abstract】 In the process of ECG acquisition, the interference of various noises will cause signal distortion and baseline drift, which will affect the accurate judgment of cardiac signal. In this paper, an adaptive algorithm based on set empirical mode decomposition is proposed. Firstly, the intrinsic mode function(IMF) component is decomposed by ensemble empirical Mode decomposition(EEMD) for ECG signals containing noise and baseline drift. Then, the IMF components that need to be processed are selected. Finally, the low order IMF with noise is processed by adaptive window and the high order IMF with baseline drift is removed to reduce the noise. The experimental results in the MIT-BIH database show that the method based on EEMD achieves good noise reduction effect. Compared with the adaptive window method based on EMD, the average signal-to-noise ratio is increased by 1.750 7, about 13%, under the same myo-electrical noise condition. For the same baseline drift, the average baseline correction rate decreased by 0.079 5, about 14%, compared with the threshold method based on EEMD.更多还原