运动行人检测与跟踪方法研究

《运动行人检测与跟踪方法研究》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、西北工业大学硕士学位论文运动行人检测与跟踪方法研究姓名：常好丽申请学位级别：硕士专业：交通信息工程及控制指导教师：史忠科20060301诬北IZ_qk人学硕士学位埝文Abstract11hetechniqueofmovingpedestriandetectionandtrackinginurbantrafficscenesisoneofthekeytechniquesforIntelligentVehicleNavigationSystem．Andtheresearchonitwilldogreatcontributiontothedeve

2、lopmentofIntelligentVehiclesandthesafetyofcitytraffic．Toensurethesafetyandsmoothnessofthepedestrianflow,thisdissertationisfocusedonthestudyofmovingpedestriandetectionandtrackinginurbantrafficscenesbyusingthedatafromafixedCCDcamera．Firstly,thisdissertationdescribestheproces

3、sofextractthepedestriansfromthecomplexscenesandhowtotrackthepedestrians．Then，themethodofpedestrians’motionanalysisisdiscussed7Fheself-adaptivebackgroundsubtractionmethodandthemotionareasegmentationmethodwereadoptedheretosegmentthemovingpedestrians．Onthebasisofbackgrounddif

4、ferenceimage，thedynamicmulti—thresholdsmethodandthemorphologicoperationwereusedtoextractthepedestriansSothenoisesbroughtbythedifferenceimageswererestrained．Andalso，theruptureintheimagesincludingpedestrianscanbejoinedtogether．Then，thepedestrianscanbeextractedexactly．Duringt

5、heprocessofmatching，thetemplateforpedestrianmatchingandrecognizinginimageseqnencewaspresentedbyfusingseveralcharactersoftargets．AnewmethodbasedongraymodelGM(1，I)wasproposedtopredictthemotionofpedestrians．Sothedetectingareaisreducedgreatly．‘Fhecenterofthetrackingwindowwasel

6、ectedasthetrackingpointrepresentingthepedestrian．Theinfluencesofthepedestrians’selfmotionwererestrained，sotherobustnessofthetrackingmethodwasimprovedTheexperimentresultsoftrackingaresatisfying．What’smore，themethodofextractiontheparametersofpedestriantrafficwasdiscussed．Thi

7、sisnecessarytocomprehendthepedestrianslmotion．ItisthetheoreticalbasistosolvetheexistingproblemsinnavigationsystemofIntelligentVehiclesandthemonitorandcontrolsystemoftraffic．Experimentalresultsoftworealurbantrafficscenesshowedthatthismethodcoulddetectandtrackpedestriansrobu

8、stlyTheresultsindicatethatthismethodiseffective．Keywords：IntelligentVehicle，trafficsafety