基于卡爾曼濾波的視頻跟蹤技術的研究及應用.pdf

1、河北工業(yè)大學碩士學位論文基于卡爾曼濾波的視頻跟蹤技術的研究及應用姓名：王書玲申請學位級別：碩士專業(yè)：計算機應用技術指導教師：彭玉青201012基于卡爾曼濾波的視頻跟蹤技術的研究及應用iiRESEARCHIMPLEMENTONVIDEOTRAKINGTECHNOLOGYBASEDONKALMANFILTERABSTRACTThetechnologyofobjectdetectiontrackinginvideosequencesisone

2、ofthehotspotsinthefieldofcomputervision.Inthecivilianusesmilitaryithasawiderangeofapplicationssuchasrobotvisionvisualprewarningmachinenavigationtrafficmanagemedicaldiagnosisintelligentmoniting.Theyhavegreatresearchapplic

3、ationvalues.Thistopicisstillaproblemespeciallywhenthebackgroundiscomplex.Suchasposevariationsoftheobjectsignificantilluminationchangesbackgroundinterferenceocclusionproblemetc.Byreflectingtheseproblemsthemainachievements

4、areobtainedasfollows:Firstlytheresearchisfocusedonthetwomajcomponentsoftracker:targetpredictedtrackinginthisdissertation.Analyzedtheusualalgithmexplainssomesimpletheetical.Secondlyintroducedseveralmainobjectfeaturesmatch

5、ingalgithms.Thispapermainlystudiesthecolimagehistogram.basedonthat2Dhistogramareproposed.AnalyzedtheMeanShiftCamShiftalgithm.2DhistogramusedintheCamShift.TheexperimentprovedthattheimprovedoftheCamShiftalgithminthetargetr

6、otationilluminationchangespartialocclusioncolfeatureofthebackgroundsimilartocolfeatureofthetargetconditionsthanthetraditionalalgithmhavebeenimproved.Finallytosolvethisproblemofcompletelyocclusionthispaperproposestheuseof

7、Kalmanfilteringtopredictthetarget.ThiskindofalgithmusetheKalmanfilteringpredictthepossiblepositiontargetoccursinthecurrentframethenusetheimprovedoftheCamShiftalgithmtotracking.Theresultsprovedthatthecombinationalgithmhas

8、excellenttrackingresultintargettotallyoccluded.Thealgithmfmovingobjecttostayalongtimespeedoftrackingtherapidlychangingcircumstanceshaveverygoodpractical.Tofurtherimprovethevideotrackingrobustnessaccuracystabilityunderthe