广西大学学报（自然科学版）
廣西大學學報（自然科學版）
엄서대학학보（자연과학판）
JOURNAL OF GUANGXI UNIVERSITY （NATURAL SCIENCE EDITION）
2014年
3期
607-614
,共8页
声发射%压缩感知%匹配追踪算法%信号重构
聲髮射%壓縮感知%匹配追蹤算法%信號重構
성발사%압축감지%필배추종산법%신호중구
acoustic emission%compressed sensing%orthogonal matching pursuit%signal reconstruc-tion
为解决声发射检测中信号采集时间长、信号量大、存储成本高的问题,提出了一种应用压缩感知理论( CS)的声发射信号采集方法,实现了应用匹配追踪算法( MP)和正交匹配追踪算法( OMP)对声发射信号进行重构,利用Matlab软件对原信号和重构信号进行参数化分析,同时对比MP算法和OMP算法对原信号的重构效果。实验结果表明：应用CS能够实现远低于传统Nyquist采样定律所要求的采样频率进行信号采样,同时实现数据的压缩；MP算法和OMP算法都能实现声发射信号的重构恢复,且OMP算法比MP算法恢复效果更好；对信号进行参数化分析,OMP算法对原信号的平均重构误差为5.56%,重构信号参数误差不到1%,表明通过压缩采样约30%的数据便可获取原声发射信号的主要信息。重构信号降低了采样率,节省存储空间,且信号不失真,CS理论在声发射无损检测领域有着广阔的应用前景。
為解決聲髮射檢測中信號採集時間長、信號量大、存儲成本高的問題,提齣瞭一種應用壓縮感知理論( CS)的聲髮射信號採集方法,實現瞭應用匹配追蹤算法( MP)和正交匹配追蹤算法( OMP)對聲髮射信號進行重構,利用Matlab軟件對原信號和重構信號進行參數化分析,同時對比MP算法和OMP算法對原信號的重構效果。實驗結果錶明：應用CS能夠實現遠低于傳統Nyquist採樣定律所要求的採樣頻率進行信號採樣,同時實現數據的壓縮；MP算法和OMP算法都能實現聲髮射信號的重構恢複,且OMP算法比MP算法恢複效果更好；對信號進行參數化分析,OMP算法對原信號的平均重構誤差為5.56%,重構信號參數誤差不到1%,錶明通過壓縮採樣約30%的數據便可穫取原聲髮射信號的主要信息。重構信號降低瞭採樣率,節省存儲空間,且信號不失真,CS理論在聲髮射無損檢測領域有著廣闊的應用前景。
위해결성발사검측중신호채집시간장、신호량대、존저성본고적문제,제출료일충응용압축감지이론( CS)적성발사신호채집방법,실현료응용필배추종산법( MP)화정교필배추종산법( OMP)대성발사신호진행중구,이용Matlab연건대원신호화중구신호진행삼수화분석,동시대비MP산법화OMP산법대원신호적중구효과。실험결과표명：응용CS능구실현원저우전통Nyquist채양정률소요구적채양빈솔진행신호채양,동시실현수거적압축；MP산법화OMP산법도능실현성발사신호적중구회복,차OMP산법비MP산법회복효과경호；대신호진행삼수화분석,OMP산법대원신호적평균중구오차위5.56%,중구신호삼수오차불도1%,표명통과압축채양약30%적수거편가획취원성발사신호적주요신식。중구신호강저료채양솔,절성존저공간,차신호불실진,CS이론재성발사무손검측영역유착엄활적응용전경。
In order to solve the problem of long acquisition time, large quantity of data, and high cost in acoustic emission signal acquisition, a theory of compressed sensing(CS) applied acoustic emission signal acquisition method was developed. The matching pursuit algorithm ( MP) and or-thogonal matching pursuit algorithm ( OMP) to reconstruct the acoustic emission signal were applied to realize parametric analysis of the original signal and the reconstructed signal by using Matlab soft-ware,and to compare reconstruction effect of MP and OMP on the original signal. The experimental results show that applying CS can achieve far lower sampling frequency than conventional Nyquist sampling theorem sampling and data compression. MP and OMP can reconstruct acoustic emission signal,but the OMP is better than MP on recovery effect. Signal parametric analysis show that the average reconstruction error is 5.56% by the OMP, signal reconstruction parametric error is less than 1% and the main information of the original acoustic emission signal can be obtain by com-pressing about 30% of sampling data. Reconstruction signal reduces signal sampling rate and saves storage spacewithout signal distortion, the CS has broad application prospects on acoustic emission nondestructive testing field.
