People Tracking in Outdoor Environment Using Kalman Filter.pdf

《People Tracking in Outdoor Environment Using Kalman Filter.pdf》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、2012ThirdInternationalConferenceonIntelligentSystemsModellingandSimulationPeopleTrackinginOutdoorEnvironmentUsingKalmanFilterMahmoudMirabiShahramJavadiElectricalEngineeringDepartmentAssistantProfessor,ElectricalDepartmentIslamicAZADUniversity,CentralTehranBranch

2、IslamicAzadUniversity,CentralTehranBranchTehran,IranTehran,Iranmahmoud.mirabi@gmail.comsh.javadi@iauctb.ac.irAbstract—Humantrackingindynamicsceneshasbeenancomputervision.Therearealotofdifficultiesforasingleimportanttopicofresearch.Objectsintheworldexhibitobjectt

3、rackinglikeilluminationvariability,backgroundcomplexinteractionswhencapturedinavideosequence,somenoiseandocclusions.Multipleobjecttrackingisevenmoreinteractionsmanifestthemselvesasocclusions.Avisualchallengingduetomultiobjectocclusions.trackingsystemmustbeableto

4、trackobjectswhichareMethodstosolvetheocclusionprobleminmultiplepartiallyorevenfullyoccluded.Inthispaper,weproposeaninteractingobjectstrackinghavebeenpreviouslypresented.approachfortrackingmultipleobjectsindynamicscenestoShiloh[6],Chang[7]andDockstader[8]overcame

5、handleobjectspartialocclusion.ThealgorithmconsistsoftwoocclusioninmultipleobjectstrackingbyusedfusingstepsatfirststepweuseGaussianMixtureModelasanmultiplecamerainputs.Cucchiara[9]proposedprobabilisticeffectivewaytoextractmovingobjectsfromavideosequence.masksanda

6、ppearancemodelstocopewithfrequentshapeThenacombinationalmethodisusedforshadowremoval.Thechangesandlargeocclusions.Eng[10]developedasecondstepoftheproposedalgorithmisobjecttrackingBayesiansegmentationapproachthatfusedaregion-basedframeworkbasedonKalmanfilteringwh

7、ichusesStableMarriageProblem(SMP)implementedalgorithmwhichisbackgroundsubtractionandahumanshapemodelforpeopleadaptedtoperformdataassociationandocclusiondetection,trackingunderocclusion.andfastmeanshiftcomputationresultsfortrackingduringThispaperpresentsamethodfo

8、rtrackingmovingocclusion.Ourapproachhastheadvantagesoflowcostandobjectsinanoutdoorenvironment.WeusedGaussianlowcomplexity,andcanberealizedinrealtimesystemandisMixture