合肥工业大学学报（自然科学版）
閤肥工業大學學報（自然科學版）
합비공업대학학보（자연과학판）
JOURNAL OF HEFEI UNIVERSITY OF TECHNOLOGY(NATURAL SCIENCE)
2014年
4期
457-461
,共5页
高压水射流%小波包%模极大值%支持向量机%最优频率范围
高壓水射流%小波包%模極大值%支持嚮量機%最優頻率範圍
고압수사류%소파포%모겁대치%지지향량궤%최우빈솔범위
high pressure water jet%wavelet packet%modulus maximum%support vector machine%op-timal frequency range
为了减少高压水射流冲击靶物的反射声信号的数据处理量，提高识别靶物材质和几何尺寸效率，文章对高压水射流反射声信号进行优化处理，利用小波变换理论进行降噪，应用小波包算法将信号按频率分层；利用模极大值算法处理各层信号，寻找能够确定靶物边界点的信号分量，确定用于几何尺寸判定的反射声信号最优频率范围；利用各层信号的能量分布作为特征值，结合支持向量机算法识别靶物材质，以最好识别率为优化指标，确定材质判断的最优频率范围的信号分量。介绍了上述算法的基本原理，编制了相应数据处理程序，利用高压水射流装置与数据采集卡设计实验装置进行实验，采集数据进行处理；确定判别靶物尺寸的反射声最优频率段为0～2.5 kHz ，判别靶物材质的最优频率段为5～20 kHz ；最终确定在特定频率段进行尺寸和材质识别，相对选用全部数据能减小数据处理量，有更高的识别效率。
為瞭減少高壓水射流遲擊靶物的反射聲信號的數據處理量，提高識彆靶物材質和幾何呎吋效率，文章對高壓水射流反射聲信號進行優化處理，利用小波變換理論進行降譟，應用小波包算法將信號按頻率分層；利用模極大值算法處理各層信號，尋找能夠確定靶物邊界點的信號分量，確定用于幾何呎吋判定的反射聲信號最優頻率範圍；利用各層信號的能量分佈作為特徵值，結閤支持嚮量機算法識彆靶物材質，以最好識彆率為優化指標，確定材質判斷的最優頻率範圍的信號分量。介紹瞭上述算法的基本原理，編製瞭相應數據處理程序，利用高壓水射流裝置與數據採集卡設計實驗裝置進行實驗，採集數據進行處理；確定判彆靶物呎吋的反射聲最優頻率段為0～2.5 kHz ，判彆靶物材質的最優頻率段為5～20 kHz ；最終確定在特定頻率段進行呎吋和材質識彆，相對選用全部數據能減小數據處理量，有更高的識彆效率。
위료감소고압수사류충격파물적반사성신호적수거처리량，제고식별파물재질화궤하척촌효솔，문장대고압수사류반사성신호진행우화처리，이용소파변환이론진행강조，응용소파포산법장신호안빈솔분층；이용모겁대치산법처리각층신호，심조능구학정파물변계점적신호분량，학정용우궤하척촌판정적반사성신호최우빈솔범위；이용각층신호적능량분포작위특정치，결합지지향량궤산법식별파물재질，이최호식별솔위우화지표，학정재질판단적최우빈솔범위적신호분량。개소료상술산법적기본원리，편제료상응수거처리정서，이용고압수사류장치여수거채집잡설계실험장치진행실험，채집수거진행처리；학정판별파물척촌적반사성최우빈솔단위0～2.5 kHz ，판별파물재질적최우빈솔단위5～20 kHz ；최종학정재특정빈솔단진행척촌화재질식별，상대선용전부수거능감소수거처리량，유경고적식별효솔。
In order to reduce the processing data volume of the reflection signal produced by high pressure water jet and to improve the identification efficiency of target material and geometric size ,the reflection signal was processed to reduce noise by using wavelet transform theory and decomposed by using wavelet packet algo-rithm .The signal component decomposed was processed by using modulus maxima algorithm and the signal component with optimal frequency range was selected which could be used to confirm the target boundary point and identify the geometric size of target .The energy distribution of different signal components was a-dopted as the eigenvalue of target material ,which was input into support vector machine algorithm to recog-nize the target material .The best recognition rate was used as the optimization index for determining the sig-nal component with optimal frequency range .The detailed basic principle of the above algorithm was intro-duced and the corresponding data processing software was programmed . The experimental equipment was built by using the high pressure water jet device and high-speed data acquisition card .The experiments were done and the acquired data was processed .The results indicate that the optimal frequency range of reflection sound signal to identify target size is 0-2.5 kHz ,the optimal frequency range of reflection sound signal to rec-ognize target material is 5-20 kHz .So the specific frequencies to identify target material and size are deter-mined ,w hich can reduce the amount of data processing and improve the identification efficiency .
