基于提升小波的信號檢測與圖像識別方法研究.pdf

1、西南交通大學(xué)碩士學(xué)位論文基于提升小波的信號檢測與圖像識別方法研究姓名：尹明申請學(xué)位級別：碩士專業(yè)：信號與信息處理指導(dǎo)教師：尹忠科20061201西南交通大學(xué)碩士研究生學(xué)位論文第11頁AbstractDifferentfromtraditionalMaUatconstructmethodWaveletliftingschemeiswhollybasedOiltime—spatialdomainItreliesonneithertransl

2、ationanddilationnorspectrmnanalyficaltoolsKeepingbiorthogonalconditiontheperformanceofwaveletanddualwaveletisimprovedthroughliftinganddualliftingYetliftingprocessofwaveletcan’tonhancevanishingmomentofwaveletThroughtheimp

3、rovementofliftingcoefficients，thethesisenhancesvanishingmomentanddualvanishingmomentaccordinglywaveletpossessesmoreconcussivecharacterInordertodetectrailwayfrequencyshiftsignalaccordingtoitscharactersamethodbased011lifti

4、ngwaveletandFouriertransformisproposedForthehomemade4informationandUM7lrailwayshiftsignal，liftingwavelettransformanditscontrarytransformaieappliedThroughdintofpeakvaine，theabscissaofsignalandfrequencydiscontinuouspointsa

5、leresolvedThenthroughDFTofsignalbetweentwodiscontinuouspointsalltheparametersofrailwayfrequencyshiftsignalareresolvedThedifferenteffectofdetectiononrailwayfrequencyshiftsignalbyHaarandbiorthogonal9／7waveletisalsocompared

6、。Offlinehandwrittensignatureverification(HSV)ishardbecauseitdepends011onlystaticinformationofimageanddynamicinformationinwritingprocessnearlyfullydisappearedFacingthecharactersofOffl／heHSVthisthesisproposesamethodbasedon

7、liftingwavelettransformItcombinesstructurefeaturewimstatisticalfeamre，thenexWaetedeigenvectoriscompressedbyKLtransformAtlasttruesignatureandforgesignaturearedistinguishedthroughsupportvectormachines(SVM)Theexperimentalre

8、sultsdemonstratethealgorithmpresentedherereachessatisfiedidenticalratefortestingstylebookLiftingwavelettransformistheleadingfocusofthewholethesis111cexcellentperformanceofliftingwavelettransformononedimensionsigual(railw

9、ayfTequencyshiftsignal)andtwodimensionsignal(image)indicatesthesignificanceofliftingwaveletinsignalprocessdomainMoreoverenhancedvanishingmomentwaveletfilterdesignedbasedonliftingschemeplaystheimportantroleinsignalprocess