首页 | 本学科首页   官方微博 | 高级检索  
     

一种在时级(频)域自动再生最佳特征参数的方法
引用本文:黄明辉,虞仲龙.一种在时级(频)域自动再生最佳特征参数的方法[J].振动.测试与诊断,1999,19(4):364-367.
作者姓名:黄明辉  虞仲龙
作者单位:中南工业大学机电工程学院!长沙410083
摘    要:针对付氏分析对时变信号不敏感和缺乏空间局部性的弱点,提出了一种基于正交小波变换的原始特征参数的构造方法,在此基础上通过遗传算法提取蕴涵于小波系数序列的故障特征,获得较理想的特征参数。实验证明,本文提出的这种方法用于特征提取与模式识别是有效的。

关 键 词:小波分析  遗传算法  模式识别  特征参数  时频域

A Method for Automatic Regeneration of Optimum Feature Parameters in Time-Frequency Domain
Huang Minhui,Yu Zhonglong,Zhang Zhiyong.A Method for Automatic Regeneration of Optimum Feature Parameters in Time-Frequency Domain[J].Journal of Vibration,Measurement & Diagnosis,1999,19(4):364-367.
Authors:Huang Minhui  Yu Zhonglong  Zhang Zhiyong
Abstract:Directed at the w eakness of Fourier analysis in being insensitive to tim e-varying signaland deficient in space partiality, and based on orthogonalw avelet decom position, anapproach to the construction of prim itive feature param eters is suggested in the paper. On the basis of the approach, failure feature included in w avelet coefficients sequence can be extracted by m eans ofgenetic algorithm and an optim um feature param etercan be obtained. It is proved by experim ents that the approach is effective in feature extraction and pattern recognition.
Keywords:waveletanalysis  genetic algorithm  pattern recognition  feature param eters  tim e-frequency dom ai
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号