基于变分模态分解的心电图信号半盲提取算法

Semi-Blind Extraction of ECG Signals Based onV ariational Modal Decomposition

针对传统心电图信号提取算法存在信号降噪效果不理想、效率低且误差较大问题,提出心电图信号半盲提取算法。采用变分模态分解方法分解输入心电图信号,获取有限个固有模态分量,按照大小顺序排列固有模态分量信号频率,同时对高频固有模态分量进行均值滤波;重构高频固有模态分量以及保留下的低频固有模态分量,结合半盲反卷积单源分离以及心电消除算法,完成心电图信号半盲提取。为验证所提方法的有效性,设计实验。实验结果表明,研究的算法降噪效果更好,能够以更高效率、更低误差完成心电图信号半盲提取,能够为相关领域研究提供可靠基础。

Due to poor denoising effect, low efficiency and big error in traditional algorithms, an algorithm semiblind extraction for ECG signal was proposed. The variational mode decomposition method was used to decompose the ECG signals and thus to obtain a finite number of intrinsic mode components. Then, the signal frequencies of intrinsic mode components were arranged in order. Meanwhile, the mean filtering was performed on the high-frequency intrinsic mode components. And then, the high-frequency intrinsic mode component and the reserved low-frequency intrinsic mode component were reconstructed. Moreover, the semi-blind deconvolution one-source separation and ECG elimination algorithm were combined to complete the semi-blind extraction of ECG signals. In order to verify the effectiveness of the proposed algorithm, a simulation experiment was designed. Following conclusions can be drawn from experimental results: the proposed algorithm has a better noise reduction effect. Meanwhile, it can complete the semi-blind extraction of electrocardiograph signals with higher efficiency and lower errors. This algorithm can provide a reliable basis for relevant researches.