資源描述:
《基于雙邊緣檢測(cè)的車牌識(shí)別算法》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、ComputerEngineeringandApplications計(jì)算機(jī)工程與應(yīng)用2013,49(8)169基于雙邊緣檢測(cè)的車牌識(shí)別算法1,21,21,2王磊,王瀚漓,何良華1,21,21,2WANGLei,WANGHanli,HELianghua1.同濟(jì)大學(xué)計(jì)算機(jī)科學(xué)與技術(shù)系,上海2018042.同濟(jì)大學(xué)嵌入式系統(tǒng)與服務(wù)計(jì)算教育部重點(diǎn)實(shí)驗(yàn)室,上海2000921.DepartmentofComputerScienceandTechnology,TongjiUniversity,Shanghai201804,China2.KeyLabofEmbeddedSy
2、stemandServiceComputing,TongjiUniversity,Shanghai200092,ChinaWANGLei,WANGHanli,HELianghua.Licenseplaterecognitionbasedondouble-edgedetection.ComputerEngineeringandApplications,2013,49(8):169-173.Abstract:Withthedevelopmentofintelligenttransportation,licenseplaterecognitionsystemhas
3、becomeanimportantpartofit.Licenseplaterecognitioncanbedividedintothreeprocedures,includinglicenseplatelocation,charactersegmentationandcharacterrecognition.Inordertoachieveaccuratelicenseplaterecognition,anovelapproachisproposedinthispaper.Duringthetaskoflocatinglicenseplate,thedou
4、ble-edgedetectionmethodisadoptedforpositioningthelicenseplate.Andthecombina-tionoffindingconnecteddomainsandtraditionalsegmentationofprojectionisappliedforcharactersegmentation.Regardingcharacterrecognition,threekindsofclassifiersareutilizedforimprovingclassificationaccuracyandthes
5、trategyofreclassifica-tionofconfusingcharactersisemployed,whichcanshortenthetrainingtimeandgethigheraccuracy.Experimentalresultsdem-onstratethatproposedapproachisabletoachievehighrecognitionratewithreasonablecomputationalcomplexity.Keywords:licenseplatelocation;edgedetection;charac
6、tersegmentation;connecteddomains;verticalprojection;characterrec-ognition;SupportVectorMachine(SVM)摘要:隨著智能交通的不斷發(fā)展,車牌識(shí)別系統(tǒng)已經(jīng)成為其中的重要組成部分。車牌識(shí)別分為車牌定位、字符分割以及字符識(shí)別三個(gè)部分。提出了一種新型車牌識(shí)別方法。在車牌定位方面,采用雙邊緣檢測(cè)車牌定位方法;對(duì)于字符分割則提出了尋找連通域與傳統(tǒng)投影分割相結(jié)合的方法;在字符識(shí)別上,將分類器分為三組,同時(shí)對(duì)于易混淆的字符進(jìn)行了再次分類,這種做法縮短了訓(xùn)練時(shí)間,提高了準(zhǔn)確率。實(shí)驗(yàn)結(jié)果
7、表明,所提出的方法具有識(shí)別率高和速度快等特點(diǎn)。關(guān)鍵詞:車牌定位;邊緣檢測(cè);字符分割;連通域;垂直投影;字符識(shí)別;支持向量機(jī)(SVM)文獻(xiàn)標(biāo)志碼:A中圖分類號(hào):TP391doi:10.3778/j.issn.1002-8331.1205-0134[7-8]隨著人們生活水平的提高和社會(huì)發(fā)展,汽車已經(jīng)逐漸像紋理特征的車牌定位方法。以上每一種方法,都有其走進(jìn)了千家萬戶,數(shù)量也在不斷增加。隨之而來的是汽車優(yōu)點(diǎn)和局限性。顏色分割容易受到光照影響,計(jì)算速度慢;管理問題,車牌識(shí)別系統(tǒng)作為管理汽車的一種重要手段已邊緣檢測(cè)的方法具有定位準(zhǔn)確率高,速度快,能有效抑制經(jīng)越來越受到重
8、視。車牌識(shí)別是通過對(duì)采集到的圖像進(jìn)噪音的特點(diǎn),但對(duì)車