基于ICEEMD和压缩感知理论的脉搏信号去噪  

Pulse signal de⁃noising method based on ICEEMD and compressed sensing

作　　者：高利雯 GAO Liwen(School of ZTE Communication,Xi’an Traffic Engineering Institute,Xi’an 710300,China)

机构地区：[1]西安交通工程学院中兴通信学院,陕西西安710300

出　　处：《电子设计工程》2020年第21期10-13,19,共5页Electronic Design Engineering

基　　金：国家自然科学基金(51207002);陕西省科技厅工业攻关项目(2018GY-066)。

摘　　要：针对脉搏信号易受噪声干扰的特性,提出了一种基于改进的完全集合经验模态分解ICEEMD和压缩感知理论的脉搏信号去噪方法。首先对脉搏信号进行ICEEMD分解,得到一组固有模态函数IMF分量,去掉最高频IMF分量,再应用压缩感知理论中的匹配追踪算法对剩余的IMF分量进行稀疏分解,选择最佳原子重构得到纯净脉搏信号的近似估计,从而实现脉搏信号的去噪。实验结果表明,提出的去噪方法在抑制噪声的同时,还有效地保留了脉搏信号的细节特性,去噪性能更好。Due to the characteristics of pulse signal is easy to be disturbed by noise,this paper presents a de⁃noising method based on Improved Complete Ensemble Empirical Mode Decomposition(ICEEMD)and compressed sensing theory.Firstly,the contaminated pulse signal were de⁃composed into a set of intrinsic mode function components using ICEEMD.Secondly,the IMFcomponents were sparsely decomposed using MP algorithm which belonged to compressed sensing theory,from the IMF components removed the highest frequency component.Then,the approximate estimation of pure pulse signal is obtained by choosing the best atom to reconstruction,and the de⁃noising of pulse signal can be realized.The result of experiment shows that this method can retain effectively the detailed characteristics of pulse signal when restraining noise,and has better de⁃noising effect.

关 键 词：脉搏信号 去噪 改进的完全集合经验模态分解 匹配追踪算法 

分 类 号：TN0[电子电信—物理电子学]