电子信息工程与通信工程专业英语 张雪英 机械工业出版社 reading电子档
Wavelet
1.Overview
Waveletsarefunctionsthatsatisfycertainmathematicalrequirementsandareusedtorepresentdataorotherfunctions.Thisideaisnotnew.Approximationusingsuperpositionoffunctionshasexistedsincetheearly1800'swhenJosephFourierdiscoveredthathecouldsuperposesinesandcosinestorepresentotherfunctions.However,inwaveletanalysis,thescalethatweusetolookatdataplaysaspecialrole.Waveletalgorithmsprocessdataatdifferentscalesorresolutions.Ifwelookatasignalwithalarge"window",wewouldnoticegrossfeatures.Similarly,ifwelookatasignalwithasmall"window",wewouldnoticesmallfeatures.Theresultinwaveletanalysisistoseeboththeforestandthetrees.
Thismakeswaveletsinterestinganduseful.Formanydecades,scientistshavewantedmoreappropriatefunctionsthanthesinesandcosineswhichcomprisethebasesofFourieranalysis,toapproximatechoppysignals.Bytheirdefinition,thesefunctionsarenon-local(andstretchouttoinfinity).Theythereforedoaverypoorjobinapproximatingsharpsignals.Butwithwaveletanalysis,wecanuseapproximatingfunctionsthatarecontainedneatlyinfinitedomains.Waveletsarewell-suitedforapproximatingdatawithsharpdiscontinuities.
Thewaveletanalysisprocedureistoadoptawaveletprototypefunction,calledananalyzingwaveletormotherwavelet.Temporalanalysisisperformedwithacontracted,high-frequencyversionoftheprototypewavelet,whilefrequencyanalysisisperformedwithadilated,low-frequencyversionofthesamewavelet.Becausetheoriginalsignalorfunctioncanberepresentedintermsofawaveletexpansion(usingcoefficientsinalinearcombinationofthewaveletfunctions),dataoperationscanbeperformedusingjustthecorrespondingwaveletcoefficients.Andifyoufurtherchoosethebestwaveletsadaptedtoyourdata,ortruncatethecoefficientsbelowathreshold,yourdataissparselyrepresented.Thissparsecodingmakeswaveletsanexcellenttoolinthefieldofdatacompression.
Otherappliedfieldsthataremakinguseofwaveletsincludeastronomy,acoustics,nuclearengineering,sub-bandcoding,signalandimageprocessing,neurophysiology,music,magneticresonanceimaging,speechdiscrimination,optics,fractals,earthquake-prediction,radar,humanvision,andpuremathematicsapplicationssuchassolvingpartialdifferentialequations.
2.WaveletversusFourierTransforms
Fourier'srepresentationoffunctionsasasuperpositionofsinesandcosineshasbecomeubiquitousforboththeanalyticandnumericalsolutionofdifferentialequationsandfortheanalysisandtreatmentofcommunicationsignals.Fourierandwaveletanalysishavesomeverystronglinks.
SimilaritiesbetweenFourierandWaveletTransforms
ThefastFouriertransform(FFT)andthediscretewavelettransform(DWT)arebothlinearoperationsthatgenerateadatastructurethatcontainslog2nsegmentsofvariouslengths,usuallyfillingandtransformingitintoadifferentdatavectoroflength2n.
Themathematicalpropertiesofthematricesinvolvedinthetransformsaresimilaraswell.TheinversetransformmatrixforboththeFFTandtheDWTisthetransposeoftheoriginal.Asaresult,bothtransformscanbeviewedasarotationinfunctionspacetoadifferentdomain.FortheFFT,thisnewdomaincontains
关键词:电子信息工程通信工程专业英语张雪英机械工业出版社reading 相关图书 电子与通信技术专业英语(第3版) 电子与通信技术专业... 作者:刘小芹 出版方:人民...
和意识。 参考书目张雪英、数字语音处理及MATLAB仿真」、电子工业出版社,2012-07...越多高校中的通信工程、电子信息工程、自动控制、计算机 技术与应用等专业开设这...
和意识。 参考书目张雪英、数字语音处理及MATLAB仿真」、电子工业出版社,2012-07...越多高校中的通信工程、电子信息工程、自动控制、计算机 技术与应用等专业开设这...
热门文档