摘要
提出了一种改进的广义谐波小波包分解算法,克服了传统特征提取方法的缺点,实现了信号的快速无混叠分离。通过与小波包分析、经验模态分解、广义谐波小波包分解进行比较,证明其在信号分析上的有效性和优越性。基于超声波信号频率非单一性的特点,将该算法应用到超声波信号特征提取中,实现了任意频段及任意频宽的信号特征提取,且计算量小。该算法为信号特征提取提供了一种更加精确有效的方法。
An improved generalized harmonic wavelet packet decomposition algorithm is presented in this paper. The shortage of traditional feature extraction methods is overcome, and the fast aliasing separation to signals is realized. It proved to be an effective and superiority signal analyzed method compared with wavelet packet analysis, empirical mode decomposition, and generalized harmonic wavelet packet decomposition. Based on the nonsingle frequency character of the ultrasonic signal, the algorithm is ap plied to the ultrasonic signal feature extraction. The signal features in any band and any bandwidth is extracted with less calculation. The algorithm provides a more accurate and effective method in the fieJd of signal feature extraction.
出处
《燕山大学学报》
CAS
2013年第4期358-365,共8页
Journal of Yanshan University
基金
国家自然科学基金资助项目(61077071
61071202)
河北省自然科学基金资助项目(F2011203207
F2010001312)
河北省科技计划项目(11213582)
关键词
广义谐波小波
小波包分解
经验模态分解
信号特征提取
generalized harmonic wavelet
wavelet packet decomposition
empirical mode decomposition
signal feature extraction
