視頻車輛檢測技術(shù)的應(yīng)用和發(fā)展.pdf

1、Y時3382學(xué)校代碼10)54密級：論五編號上洛絡(luò)事天擎SHANGHAlMARITIMEUNIVERSITY碩士學(xué)位論文MASTERDISSERTATION論文題目：塑塑奎翅掛型籃壟曲廛旦狃蕉星學(xué)科專業(yè)：過簋盟撞盔指導(dǎo)教師：墨基ill麴撞2瞳噩塑l叢童雖2完成日期：三QQ亟至主旦ABSTRACTABSTRACTThetrafficsurveillanceisoneoftheimportanceapplicationsinIntellig

2、entTransportationSystemItsurveystheflowofvehicletraffic；collectallparameterofthetrafficflow，forexamplecapacity，speed，vehicletype，andqueuetimeandqueueextentThesuitabletechnologiesofvehicledetectionplaytheimportantroleinth

3、etrafficsurveillanceVideovehicledetectionsystemisacomputerprocessingsystemwhichusesimageprocessingtechnologytodetectthetrafficflowItisthegoal，whichcandetect，confimltheposition，recognize，track，analysisthetrafficbehaviorof

4、vehicleandthepeoplefrOmanalyzetheimageVideovehicledetectionsystemcannotonlycollectcompletetrafficflowparameterbutalsoassistthetrafficmanagementandcontrol，andreallyachievetheintellectualizedtrafficmanagementFordecreasingi

5、nfluencetotheimagequailty，weshouldpercolatethroughtheyawpbeforetheimageprocessingVehiclereorganizationhasthreestep：imagesegmentation，vehicletrackingandvehicleclassificationThesegmentationofmovingobjectsusesamethodofframe

6、imagedifferenceBackgroundimagecanbeacquiredthroughimagesequence，makeadifferencebetweenthecurrentlyimageandbackgroundimageNextswelltheimagecontrastTrytoprocurethesuitablecontrastthresholdbyusingthemethodofthesmallestwrong

7、probabilityMovingobjectswouldbesegmentedbyanalyzingimagesbasedonthecontrastthresholdFordecreasingcountingweusematchingcharactertotracktheobjectTowardthevehicleclassification，weusethemethodofmatching3Dmodelprojectioninthe