基于分?jǐn)?shù)傅立葉變換的多分量線性調(diào)頻信號參數(shù)估計(jì)算法.pdf

1、中國科學(xué)技術(shù)大學(xué)碩士學(xué)位論文基于分?jǐn)?shù)傅立葉變換的多分量線性調(diào)頻信號參數(shù)估計(jì)算法姓名：周剛毅申請學(xué)位級別：碩士專業(yè)：信號與信息處理指導(dǎo)教師：沈鳳麟張長耀20021001AbstractAccordingtothesignalmodelofSARalongazimuthdirection，inthispaperanewparameterestimationmethodbasedonfractionalFouriertrailsfornlis

2、developedandthecorrespondingfastcomputationalgorithmisalsopresentedThephysicalconceptofthenewlydevelopedmethod，itscharacteristicsanditscurrentlimitationsarealsounderinvestigationInliterature，F(xiàn)ractionalFourierTransform(FR

3、FT)wasfirstintroducedinoptics，whichWaSusedtodescribetheu#tpropagationalongtheopticaxisbehindopticlenseItisakindoflineartransformbelongingtothefamilyoftimefrequencyanalysisItwasbloomedinthefieldofsignalprocessingaRer1995T

4、hephysicalmeaningisthatitistheprojectionofthetimefrequencydistributionofagivenfunctionontoallaxismakingananglewiththetimeaxisinthetimefrequencydomainThereforeFRFrhasacloserelationwitIlothertime—frequencyanalysismethodssu

5、chaswavelettransformandWignerVilledistributionItisquIlefitforprocessingtransientsignalsOnthesignalprocessingforSAR，themostdifficultistoprocessthedataalongazimuthdirectionAsknowntoall‘thedataalongazimuthdirectioncanbemode

6、ledasmulticomponentLinearFMsignals，whicharecomposedofmanyLinearFMswiththesamechirprateandDopplercenterfrequencybutwithdifferentinitialphaseHence，thekernelforprocessingtheazimuthdataistherealtimeestimationalgorithmfurthec

7、hirpconstantandDopplercenterfrequencyUponaboveunderstanding，inthispaperWearefocusedonpresentinganewmethodforsuchestimationpurposeThismethodconcentratestheenergyofmulticomponentLFMthroughthefollowingthreesteps：concentrati

8、ngtheenergyofLinearFMinthetimefrequencydomainviaFRFTmappingthetransformedsignaltoacertaintwodimensionaldomain，wheretheenergyisconcentratedalongaparticularline，bythenewlyconstructedmatchfilter；employingRadontransformtheen

9、ergydistributedalongtheaboveparticularlinewillbegatheredintoapointwhichisonlydominatedbythechirprateandDopplercenterfrequencyThereforeby廿ackingthepointwiththemaximumvaluethetwoparametersofthemulticomponentLFMcanbeestimat

10、edAccordingtotheabovethreesteps，itisobviousthattheproposedmethodhasagoodperformanceinprocessingsuchsignalCOrrUptedinwhitenoiseThefastcomputationalgorithmfortheproposedmethodisalsopresentedThecomputationloedofthealgerithm

11、issameasthatofFRFTandtheaccuracydependsonthatoftwoconsecutiveFRFToperationsLimitedbythecomputationalgorithmofFRFTrFRFTisacuttingedgeresearchtopic坷nowTherefore，thefastcomputationmethodofFⅪ叮hasnotbeensolvedperfectly)，thepr