基于動物聲音的個體辨認技術(shù)的研究.pdf

1、東北林業(yè)大學碩士學位論文基于動物聲音的個體辨認技術(shù)的研究姓名：黃英來申請學位級別：碩士專業(yè)：計算機應用技術(shù)指導教師：蘇健民20060601AbstractAbstractWiththenumberofwildanimalshassteadilydeclined，peoplebegintoplaceparamountimportanceontheprotectionofthemandresearchonthemisalsoincreasi

2、ng；Therefore，theresearchonanimalacousticshasdrawnpeople。SattentionTaxonomistsaretakingthesoundsignalofanimalsasastandardofclassificationofanimalthroughinvestigation；eco!ogistsareprobingintohowtevealthemysteryofanimalacti

3、onbysoundsignalTobemorenoticeable，animalsoundcanbeusedtodirectanima!sandmakethemdoashumanbeing’ScommandThispaperpresentsathoroughstudyoftheendpDintdetectionofEnergyFrequencyValue，LPCC，MFCCandtheirdifferencecharacteristic

4、parametertherecognitionmethodsofapplyingVQandHMMtosoundrecognitionbystudyingthebasicelementsandsystematicstructureofsoundrecognitionandanalyzingthepresenttechniqueofsoundrecognitionThus，thesystemofindividualidentificatio

5、nbasedonanimalsoundisestablishedThispaperobtainsthecharacteristicparameterofsoundinputbyMATLABsound，proeeRsingtoolbox，inwhichLPCC，MFCCandtheirdifferencearetakenasthecharacteristicparameterofsoundsignal，introducedLBGalgor

6、ithmgoingonyardsofbookdesign，ViterbialgorithmisadoptedtodotrainingandidentificationAhigherdiscriminationisachievedAndfurthermoreithastheadvantagesoffasteroperatingspeed，littlecalculationamount，lowererrorrateandSOonThispa

7、permakesabasicandpreparatorystudyforexploringpracticalsoundrecognitionsystemfurtherbytestingandinvestigatingactualsoundrecognitionsystemKeywords：EnergyFrequencyValue；LinearPredictionCepstrumCoefficient；MelFrequencyCepstr