欢迎来到天天文库
浏览记录
ID:36645036
大小:1.10 MB
页数:52页
时间:2019-05-13
《车牌识别系统中定位算法的研究》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、南京航空航天大学硕士学位论文车牌识别系统中定位算法的研究姓名:杨萱申请学位级别:硕士专业:通信与信息系统指导教师:曹杰20080101车牌识别系统中定位算法的研究AbstractWiththedevelopmentofinformationtechnologyandintelligencetechnology,theinformatizationandintelligentizingoftrafficmanagementisthetrend.LicensePlateRecognitionsystem(LPR)isthecoreofIntelligentTrafficSystem(ITS).
2、Itisveryimportantinmoderntrafficmanagementsystems.Asacomprehensivereal-timecomputervisionsystem,thevehicleslicenseplaterecognition(LPR)mainlyincludesthedivisionoflicenseplatelocalizationandrecognitionofcharacter.Becauseofthecomplexityofthescene,theunpredictabilityoflicensepositionandpicturequality
3、,thelicenseplatelocalizationdivisionsystemisstillsomethingunsatisfied.Thereforethelicenseplatelocalizationdivisionalgorithmisalwaysahotspotofresearchinthisfield.Ahybridlicenseplatelocationschemethatisbasedonmathematicalmorphologyispresentedinthispaper,whichismainlyimplementedasfollowingseveralstep
4、s.Atfirst,wegetthegrayimageofthelicenseplatedothetop-hatandthebottom-hattransformation,then,infirstCCA(Connectedcomponentanalysis)stage,wegetridofthosenon-textroughlyaccordingtothecharactersoflicenseplatehavethesimilarfeatureinheightandsize,andtheyalsoareconnectedwitheachother.Later,weconnectthese
5、platecharactersintoonecomponentbasedonfindingthecenterlineofthemusingHoughtransform.Furthermore,thesecondCCAstageisaccuratelyextractingthelicenseplateaccordingtherealfeatureofthelicenseplate.Basedontheexperimentsdealingwithimagestakenundervariousrealworldconditions,theresultsprovethatthisproposedm
6、ethodcanrelativelylocatelicenseplate.Itshowsthattheperformanceofthesystemispromising,robustandtimely.Keywords:licenseplatelocation,binarization,mathematicalmorphology,connected-componentanalysis,top-hattransform.II南京航空航天大学硕士论文图表清单图1.1车牌识别系统组成........................................................
7、................................5图2.1车牌定位系统流程图.................................................................................11图3.1蓝底白字车牌原图像......................................................................
此文档下载收益归作者所有