基于計(jì)算機(jī)視覺的車架號提取.pdf

1、西南科技大學(xué)碩士學(xué)位論文基于計(jì)算機(jī)視覺的車架號提取姓名：王新成申請學(xué)位級別：碩士專業(yè)：通信與信息系統(tǒng)指導(dǎo)教師：宋麗梅20080416西南科技大學(xué)碩士研究生學(xué)位論文第1I頁AbstractIntheprocessingofvehiclemanagement，itisnecessarytoVerifyVehicleIdentificationNumber(shortedasVIN)，thatis，tocomparetheVINwiththe

2、exfactoryrecordsoncontentandsizeAtpresent，VrNisusu“l(fā)ygainedbymanual11lbbings，whichisinconvenientandinefjFicientWiththerapidgrowthofautomobilepossession，thenlbbingsc鋤notmeettheneedsofhighomceef五ciencyInordertesolVetheprob

3、lem，VINextractionfromimagebasedoncomputerVisionispresentedinthisthesisOnthebaseofimageprincipleofcomputerVisionandwiththeaidoftargetplacedinthesameplanewithVIN，themappingrelationsbetweenimageplaneandspaceplaneofVINisobta

4、ined仃omsin91eimageT11len，therealmorphologyofVINcanbedetemlinedquantitatiVely，reachingtheVINnonpaperrubbingTheconcreteprocessisasf01lows：firstly；accordingto2Dtargetcalibrationmethodforvideocamera，theintrinsicparametersofv

5、ideocameraaregainedSecondly；thetargetintheimageislocatedbyaseriesofimageprocessingThirdly；featurepointsofthetargetintheimagearedetectedbyLevenbergMarquardtcircledetectiontechnologyandHa耐scomerdetectiontechn0109yFinally，m

6、ereconstructionofVrNiscompletedbytheimageconst九】ctionmethodprovidedinthisthesisTheexperimentalresultsprovethefeasibilityofthemethodproposedinthisthesisComparedwiththetraditional11lbbingmethod，thismethodmakesimageprocessi

7、ngandanalysisachievedbycomputer，reducingtheworkloadf0rworkers舳dgreatlyimprovingtheefjEiciencymeanwhile，itsupportsbatchprocessingandextractsVINindigitalimagefom，bene6tingVINelectronicmanagementandnetworktransmissionKeywor