摘要
为了在信号瞬时特征提取过程中有效降低噪声干扰影响,提出一种基于小波阈值降噪和经验模态分解(EMD)的信号瞬时特征提取方法。根据信号特征选择适合的小波阈值函数进行降噪处理,然后对降噪信号进行EMD分解,以互相关系数作为判别依据,保留含有信号瞬时特征的本征模函数(IMF),并进行Hilbert时频谱图和边际谱图分析,最终完成信号瞬时特征的提取。
In order to reduce the impact caused by noise pollution upon the signal processing of instanta-neous feature extraction,a method combining wavelet threshold de noising and empirical mode decomposition （EMD）to extract the instantaneous features of signals is proposed here.The method selects a suitable wavelet threshold which depends on the feature of signal to make pretreatment to the signal,and then the intrinsic mode functions （IMFs）of de noising signals are obtained with EMD.Basing on the correlation degrees,the real IMFs are extracted to an-alyze the Hilbert spectrum and the marginal spectrum.Finally we complete the extraction of signals instantaneous feature.
出处
《机械与电子》
2014年第10期12-16,共5页
Machinery & Electronics
基金
军队通保科研基金资助项目(装通[2012]80号)
