小波包时频分析方法的研究及应用  

Research and application of wavelet packet time-frequency analysis

作　　者：罗静[1] 钟佑明[2] 

机构地区：[1]重庆交通大学应用技术学院,重庆400074 [2]重庆交通大学信息科学与工程学院,重庆400074

出　　处：《重庆邮电大学学报（自然科学版）》2009年第3期379-387,共9页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：重庆市科学技术委员会自然科学基金资助项目(CSTC;2005BB6056);重庆市教育委员会高等教育教学改革研究项目(0834138)

摘　　要：针对非线性、非平稳信号分析与处理方法不能满足某些特别的工程要求的问题,提出将小波包变换和短时傅立叶变换融合形成小波包时频分析技术。建立了一套较完善的小波包时频分析技术框架体系。给出了小波包时频分量谱与小波包时频分量幅度谱以及小波包时频谱与小波包时频幅度谱的算法。对小波包分解的直接算法、Mallat算法、混合算法在计算量和识别精度等方面进行了比较研究。应用Visual C++,OpenGL,photoshop等计算机软件工具和虚拟仪器技术开发了一套虚拟小波包时频分析仪,能实现任意信号的小波包分解和显示。最后将小波包时频分析应用于仿真测试信号与实测机械磨床振动信号,发现小波包时频分析在识别奇异、辨别信号深层次细节方面具有较好的性质。For non-linear, non-stationary signal analysis and processing methods can not meet the specific requirements of the project, the wavelet packet transform and the short-time Fourier transform were integrated to form wavelet packet time-frequency analysis technology. An integrated system of the wavelet packet time-frequency analysis was established, and the algorithm of the wavelet packet time-frequency part spectrum and the part amplitude spectrum was presented. The calculation and identification precision of direct algorithm, Mallat algorithm and mixed algorithm were compared. A wavelet packet analysis virtual instrument was developed, which can decompose and display the wavelet packet signals and calculate the wavelet packet time-frequency part spectrum by means of computer software like VC ＋ ＋, OpenGL, Photoshop, etc. At last, the wavelet packet time-frequency analysis was applied in signal emulation test and vibration signal test of grinding machines, and the wavelet packet time-frequency analysis had excellent features in distinguishing and realizing the strange signals.

关 键 词：虚拟仪器 小波包时频分析 快速傅立叶变换 短时傅立叶变换 

分 类 号：TN911.72[电子电信—通信与信息系统]