improved object tracking with camshift algorithm

《improved object tracking with camshift algorithm》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、IMPROVEDOBJECTTRACKINGWITHCAMSHIFTALGORITHM**,***OULD-DRISNouar,GANOUNAliandCANALSRaphaël*LaboratoryofElectronics,SignalsandImages(LESI),UniversityofOrléans,FRANCE**FacultyofEngineering,EEDepartment,UniversityofGARYOUNIS,LIBYANouar.Ould-Driss@univ-orleans.fr,Ali.Ganoun@univ-orleans

2、.fr,Raphael.Canals@univ-orleans.frABSTRACTliterature.Ournewtrackingsystemisdescribedinsection3andtheobtainedresultsarepresentedinsection4.WeInthispaper,wepresentanewobjecttrackingapproachfinishbyaconclusionandproposesomeorientationsforbasedontheanalysisofatwo-dimensionalimagefuture

3、research.distributionhistogramcalculatedfromtwocolorimetricchannelsautomaticallyselectedoncriteriaof2.THECAMSHIFTALGORITHMrepresentativeness.Amongtheessentialcontributionsofthiswork,wecanquoteabettermodellingoftheobjecttoTheprincipleoftheCamShiftalgorithmisgivenin[04]andtrackandthe

4、managementofthetargetappearancechanges[08].Eachimageofthesequenceisconvertedintoaduringthesequence.OurapproachisaprolongationoftheprobabilitydistributionimagerelativelytothehistogramofCamShiftalgorithmapplications(ContinuouslyAdaptivetheobjecttobetracked.Fromthisimage,thecentreandt

5、heMeanShift)inordertotrackobjectpresentingstrongsizeoftheobjectaregiventhankstotheCamShiftmodificationsofshapeandluminosity.algorithm.Thesenewcentreandsizeareemployedtoplacethesearchwindowinthenextimage.Thisprocessisthenrepeatedforacontinuoustargettrackinginthevideosequence.1.INTRO

6、DUCTIONThealgorithmofCamShiftthusemploysa2Dprobabilitydistributionimageproducedfromaback-projectionoftheIntheliterature,manymethodshavebeendevelopedtotargethistogramwiththeimagetoprocess.solvetheobjecttrackingproblem.NumerousapproachesTheCamShiftalgorithmcallsupontheMeanShiftonetoa

7、rebasedonthevisualprimitivesdetectedandtrackedincalculatethetargetcentreintheprobabilitydistributionimagesbyemployingcorrelation.Othertechniquesprocessimage[9].Itisamatteroffindingarectanglepresentingthetheedgeorregiondata,orthemovementoftheobjectinsamemomentsasthosemeasuredonthepr

8、obabilityimage.ordertotrac