论文部分内容阅读
本文将集合经验模态分解(EEMD)与小波软阈值去噪算法相结合,提出了一种新的心电图信号去噪EEMD-WS算法.算法首先对信号进行EEMD分解得到有限个固有模态函数(IMF);其次,根据实际含噪心电信号中各成分的特性,将所有IMF分为低阶含噪、中阶有用信号和高阶含基线漂移三类,对于低阶含噪IMF利用IMF能量变化分界点自适应地确定含噪IMF个数,随后对其利用小波收缩算法中的启发式软阈值选择算法进行去噪;对于高阶含基线漂移IMF根据其自身是否包含周期信息自适应地判断并去除与基线漂移关系密切的IMF.最后通过将滤除噪声的低阶IMF、中阶有用信号重构达到抑制噪声和去除基线漂移的目的.仿真信号和MIT-BIH心电数据库真实心电信号实验显示,EEMD-WS算法不仅能够克服小波去噪算法不能去除基线漂移的不足,而且能够比常用的EMD-WS算法更好地提高消噪效果,总体去噪性能优于传统算法.
In this paper, we combine EEMD with wavelet soft threshold denoising algorithm, and propose a new EEMD-WS algorithm for ECG signal denoising.Firstly, the signal is decomposed by EEMD to obtain a finite number of intrinsic mode functions IMF). Secondly, according to the characteristics of each component in the actual noisy ECG, all the IMFs are classified into low-order noisy, medium-order useful signals and high-order baseline drift. For low-order noisy IMFs, The change demarcation point adaptively determines the number of noisy IMFs, and then uses the heuristic soft threshold selection algorithm in wavelet shrinkage algorithm to denoise it. For high-order IMFs with baseline drift, they are adaptively judged according to whether they contain periodic information And remove the IMF which is closely related to the baseline drift.Finally, by reducing the low-order IMF and the intermediate-order useful signal, the noise suppression and baseline drift removal are achieved.The simulation signal and the real ECG signal of MIT-BIH ECG Experiments show that the EEMD-WS algorithm can not only overcome the shortcoming that wavelet denoising algorithm can not remove the baseline drift, but also can improve the noise reduction effect better than the commonly used EMD-WS algorithm, and the overall denoising performance Better than traditional algorithms.