Hilbert-Huang变换的鼾音信号谱分析  

Snoring signal spectrum analysis based on Hilbert-Huang transform

作　　者：张引红[1] 李全禄[1] 

机构地区：[1]陕西师范大学物理学与信息技术学院,西安710062

出　　处：《计算机工程与应用》2011年第8期11-13,17,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.10374064);中央高校基本科研业务费专项资金(No.2010ZYGX020)~~

摘　　要：在分析Hilbert-Huang变换算法的基础上,利用此变换对打鼾者的鼾音信号进行了分析。通过经验模态分解把鼾音信号分解为一系列固有模态函数,并分析了各固有模态的频率特征,对各模态的生物学意义进行了描述。对固有模态函数进行了Hil-bert变换,建立了鼾音信号的Hilbert谱和边际谱。结果表明Hilbert比小波变换所建立的时频分布具有更好的时频分辨率,解决了时间分辨率和频率分辨率互相影响的问题;从实际看边际谱比傅里叶谱有更准确的物理意义。Hilbert谱和边际谱为脉搏信号的特征提取和模式识别提供了可靠的依据。Based on Hilbert-Huang Transform（HHT）,the snore signal is analyzed.A serial of Intrinsic Mode Functions（IMF） are obtained by Empirical Mode Decomposition（EMD）.The frequency of each IMF is analyzed and meaning of the biology is described.The Hilbert spectrum and the marginal spectrum of snore signal are established by HHT.The results show that Hilbert spectrum has higher time-frequency resolution than the time-frequency distribution established by wavelet transform and the interaction of the time resolution and the frequency resolution is solved,besides the marginal spectrum has more precise physical meaning than Fourier spectrum.So,HHT provides reliable basis for the feature extraction and pattern recognition of snore signals.

关 键 词：希尔伯特-黄(Hilbert-Huang)变换 鼾音 时频分布 信号分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]