欢迎来到天天文库
浏览记录
ID:36353373
大小:4.68 MB
页数:94页
时间:2019-05-09
《基于模式识别的图像处理及其在车牌识别中的应用》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、合肥工业大学硕士学位论文基于模式识别的图像处理及其在车牌识别中的应用姓名:姜滔申请学位级别:硕士专业:检测技术及自动化装置指导教师:王建平2002.12.1ImageprocessingbasedonpatternrecognitionanditsapplicationtolicenseplaterecognitionAbstractImageprocessingbasedonpaternrecognitionisoneofthemostimportantresearchdirectionsintheimagerecogniti
2、onfield.LicenseplateRecognitionbasedonimageisanimportantapplicationincomputervisionandpaternrecognitionoftheintelligenttransportationfield.Anditisalsothecoretechnologyintheintelligenttransportationsystem.Themaincontentofthispaperisthestudyonimageprocessingbasedonpate
3、rnrecognitionanditsapplicationinlicenseplatereco娜tionsystem.Inthisarticle,paternrecognitionisusedastheoryfoundationandimageprocesstechnologyasbasictechnology.Combiningwiththenewtheorytools(waveletanalysisandneuralnetwork),thispaperpresentsefectiveresolvemethodstothre
4、epartsofimagerecognition(imagesegmentation,imagefeatureextractionandimageclassificaton)thoughtsystemicanalysisofit.Inthisarticle,theresearchofimageprocessbasedonpaternrecognitionconsistsofthreephases:1)Phaseofimagesegmentation:amethodofmulti-layerandmulti-character's
5、segmentationintransformfieldbasedonknowledgeisproposed.Appliedinlicenseplatesegmentationproblem,anewsegmentationmethodofautomobilelicenseplatebasedonwavelettransformandneuralnetworkispointedout1711.2)Phaseofimagefeatureextraction:combinedwiththefeatureextractionofstr
6、ucturalandstatisticalmethod,amethodofimagecharacterfeatureextractionbasedonwaveletandmomentsanalysisispresentedt701.3)Phaseofimageclassificaton[731:afterinvestigationonintelligencerecognitiontechnology,thepaperputsforwardbasicstructureofrecognitionmachine'smodel,andm
7、akesaprimaryresearchofbasicstructureanddesignmethod,thenmakesresearchofthemulti-charactermethod.Theimageprocessmethodofthispaperhassomereferencevalueinotherimageprocessfields.Keyword:paternrecognitionimageprocessvehicle'splaterecognitionwavelettransformneuralnetwork独
8、创性声明本人声明所呈交的学位论文是本人在导师指导下进行的研究工作及取得的研究成果.据我所知,除了文中特别加以标注和致谢的地方外,论文中不包含其他人已经发表或撰写过的研究成果,也不包含为获得一一O/J匹逃一或其他教育机构的学位或证书而使用过的材料。与我一同工作的同
此文档下载收益归作者所有