• 首页
  • 期刊简介
  • 编委会
  • 投稿指南
  • 收录情况
  • 杂志订阅
  • 联系我们
引用本文:赵斯琪,彭钰莹.基于改进集合经验模态分解的信号处理方法研究[J].软件工程,2024,(8):66-69.【点击复制】
【打印本页】   【下载PDF全文】   【查看/发表评论】  【下载PDF阅读器】  
←前一篇|后一篇→ 过刊浏览
分享到: 微信 更多
基于改进集合经验模态分解的信号处理方法研究
赵斯琪, 彭钰莹
(大连东软信息学院软件与大数据技术系, 辽宁 大连 116023)
zhaosiqi@neusoft.edu.cn; pengyuying@neusoft.edu.cn
摘 要: 针对信号通过集合经验模态分解(Ensemble Empirical Mode Decomposition,EEMD)处理后获取的缺陷特征不明显的问题,提出一种EEMD联合主成分分析法(Principal Component Analysis,PCA)的信号处理方法。此方法首先对信号进行EEMD处理;其次在重构信号时,采用相关系数法进行特征选取;最后采用PCA对缺陷信号分量进行处理,能够在有效抑制模态混叠现象的同时,加强信号特征。通过仿真验证得到,经改进后的信号处理方法的重构信号信噪比为29.523 7dB,相较于单一EEMD方法提高了51.06%,均方根误差降低了51.02%。改进后的算法能够在去除噪声的同时,保留更多的信号特征,更适用于处理超声检测信号。
关键词: 信号处理;EEMD;PCA
中图分类号: TP391.9    文献标识码: A
基金项目: 辽宁省教育厅青年项目(LJKQZ20222450)
Research on Signal Processing Method Based on Improved Ensemble Empirical Mode Decomposition
ZHAO Siqi, PENG Yuying
(Department of So f tware and Big Data Technology, Dalian Neuso f t University of In f ormation, Dalian 116023, China)
zhaosiqi@neusoft.edu.cn; pengyuying@neusoft.edu.cn
Abstract: Aiming at the indistinct defect features obtained after signal processing through Ensemble Empirical Mode Decomposition (EEMD), this paper proposes a signal processing method that combines EEMD with Principal Component Analysis ( PCA). With this method, the signal is first processed by EEMD, and then during signal reconstruction, feature selection is performed using the correlation coefficient method. Finally, PCA is applied to process the defective signal components, effectively suppressing mode mixing phenomena while enhancing signal features. Simulation results show that the reconstructed signal-to-noise ratio of the improved signal processing method is 29.5237dB, an increase of 51.06% compared to the single EEMD method, and the root mean square error reduces by 51.02% . The improved algorithm can eliminate noise while preserving more signal features, making it more suitable for processing ultrasonic detection signals.
Keywords: signal processing; EEMD; PCA


版权所有:软件工程杂志社
地址:辽宁省沈阳市浑南区新秀街2号 邮政编码:110179
电话:0411-84767887 传真:0411-84835089 Email:semagazine@neusoft.edu.cn
备案号:辽ICP备17007376号-1
技术支持:北京勤云科技发展有限公司

用微信扫一扫

用微信扫一扫