計算機(jī)視覺和模式識別在車牌識別中的應(yīng)用

1、華中科技大學(xué)碩士學(xué)位論文計算機(jī)視覺和模式識別在車牌識別中的應(yīng)用姓名：沈世旻申請學(xué)位級別：碩士專業(yè)：控制理論與控制工程指導(dǎo)教師：盛翊智20040510華中科技大學(xué)碩士學(xué)位論文AbstractLicensePlateRecognitiontechniqueshavebeenwidelyusedinmodemtransportationandplayanactiveroleinincreasingproductivityandreducing

2、laborintensityThisstudyisbasedonthe“LicensePlateAutoRecognitionSystem”projectItinvestigatesrelatedtheoriesofcomputervisionandpatternrecognition，designsalicenseplaterecognitionalgorithm，andappliesteal—worldscenariosTheaut

3、omaticlicenseplaterecognitionisrealizedinthisstudyThebasiclicenseplaterecognitioncallbeclassifiedintotwocategories：blackandwhite(grayscale)imageandcolorimageIncontrasttoblackandwhiteimages，colorimagesprovidemoreinformati

4、onandbetteradaptationtosurroundingenvironmentsMultiplepossiblecombinationsofthecoloroflicenseplatesuchaswhitewordsonbluebackground，blackwordsonyellowbackground，blackwordsonwhitebackgroundandSOOnnecessitiestheanalysisofva

5、rioussituationsHoweverforin—doorlicenseplaterecognition，satisfactoryresultsCanbeobtainedbyanalyzingonlyblackandwhitefigurebecauseofsimilarenvironmentsThisresearchfocusesonblackandwhiteimages，reducesthecomputationalcostof

6、classicalalgorithmsforthetransformationofRGBtoHLSwithanewfastconversionalgorithm，andmeetsthereal—timerequirementsofthesystemMostoftheclassicalimagesegmentationmethodsutilizebinarizationtodifferentiateobjectanditsbackgrou

7、ndHoweverthismethodrelieslargelyonthecharacteristicsofboththeobjectandbackground，andthequalityofsegmentationiseasilyimpactedbytheenvironmentThisstudyproposesbinarizationafteredgedetecting，whichcallaccuratelyseparatelicen

8、seplatefromitsbackgroundandserveasthebasisforlocatingandrecognizingindividualnumbersandlettersOnlicenseplatesAttherecognitionstage，weproposethecombinationofcharacteristicinputandMulti—layerPerceptionNetwork(MLPN)，whichbo