首页 | 官方网站   微博 | 高级检索  
     

脉冲漏磁信号的EMD小波阈值去噪研究
引用本文:张韬,左宪章,田贵云,张云,费骏骉.脉冲漏磁信号的EMD小波阈值去噪研究[J].计算机工程,2012,38(7):263-265.
作者姓名:张韬  左宪章  田贵云  张云  费骏骉
作者单位:1. 军械工程学院电气工程系,石家庄050003;73681部队,安徽黄山245000
2. 军械工程学院电气工程系,石家庄,050003
3. 纽卡斯尔大学电气电子与计算机工程学院,纽卡斯尔NE1 7RU
基金项目:河北省自然科学基金资助项目(E2008001258)
摘    要:为有效抑制脉冲漏磁检测信号中的各种噪声,将小波阈值去噪运用到经验模态分解(EMD)中,提出一种基于EMD的小波阈值去噪方法。针对小波软、硬阈值函数中存在的不足,引入平滑截断绝对偏差惩罚因子进行优化改进。将该方法应用于脉冲漏磁信号进行实际消噪处理。实验结果表明,该方法能较好地剔除信号中的噪声,在脉冲噪声的抑制方面优于小波阈值去噪。

关 键 词:经验模态分解  去噪  阈值  平滑截断绝对偏差  脉冲漏磁检测
收稿时间:2011-05-23

Research on EMD Wavelet Threshold Denoising of Pulse Magnetic Flux Leakage Signal
ZHANG Tao , ZUO Xian-zhang , TIAN Gui-yun , ZHANG Yun , FEI Jun-biao.Research on EMD Wavelet Threshold Denoising of Pulse Magnetic Flux Leakage Signal[J].Computer Engineering,2012,38(7):263-265.
Authors:ZHANG Tao  ZUO Xian-zhang  TIAN Gui-yun  ZHANG Yun  FEI Jun-biao
Affiliation:1(1.Department of Electrical Engineering,Ordnance Engineering College,Shijiazhuang 050003,China;2.School of Electrical Electronic and Computer Engineering,University of Newcastle,Newcastle NE1 7RU,UK;3.73681 Forces,Huangshan 245000,China)
Abstract:In order to defect information accurately from the Pulse Magnetic Flux Leakage(PMFL) signal,a new wavelet threshold denoising method is presented based on Empirical Mode Decomposition(EMD).The Smoothly Clipped Absolute Deviation(SCAD) rule is introduced to optimize the conventional wavelet soft and hard threshold function.The new denoising method is applied to denoise the Pulse Magnetic Flux Leakage(PMFL) signal.Experimental results show that not only the new denoising method can remove the noise in signal,but also the denoising performance for pulse noise is better than conventional wavelet threshold denoising method.
Keywords:Empirical Mode Decomposition(EMD)  denoising  threshold  Smoothly Clipped Absolute Deviation(SCAD)  Pulse Magnetic Flux Leakage(PMFL) test
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《计算机工程》浏览原始摘要信息
点击此处可从《计算机工程》下载全文
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-23

京公网安备 11010802026262号