視頻中移動對象跟蹤技術(shù)研究【畢業(yè)論文】

1、本科畢業(yè)設(shè)計本科畢業(yè)設(shè)計（2020屆）屆）視頻中移動對象跟蹤技術(shù)研究所在學(xué)院專業(yè)班級計算機科學(xué)與技術(shù)學(xué)生姓名學(xué)號指導(dǎo)教師職稱完成日期年月IIAbstract【ABSTRACTABSTRACT】Withthedevelopmentofcomputertechnologycomputingpowerhasbeengreatlyimprovedmakinguseofcomputertorealizeofhumanvisualfunctionh

2、asbecomeoneofthemostpopulartopicsincomputerscience.Trackingmovingobjectsinthevideo(objecttracking)isabranchofcomputervisionwhichhasalotofsignificantapplicationsinthevideosurveillancehumanmachineinterfacesoon.Thesocalledo

3、bjecttrackingreferstothedetectionextractiontrackingofmovingtargetsinimagesequences.thenobtainthemotionparameterssuchastrajectypositionvelocityaccelerationsoon.Asahotareaofcomputervisionmanyscholarsathomeabroadproposedlot

4、sofsolutions.Howevertheseproposedmethodsstillhavetheirownspecificapplicationsthetrackingtechnologyisnotyetfullyintegrated.ThispaperpresentsacombinationofmeanshiftKalmanfilterwithupdatetrackingalgithmfmovingobjects.Firstl

5、yaccdingtothenatureofthemovementofobjectstoconstructtheKalmanfiltermodelthenusingittopredictthenextframeofthegoalscenterasthemeanshiftalgithmsinitialiterationvalue.Thenusingthecidate’scentercalculatedbymeanshifttogettheB

6、hattayyacoefficientbetweencidatetargetmodel.AccdingtothecaseofBhattayyacoefficienttoupdatetheKalmanfilter’sconversionmeasurementnoisecovariancematrix.Withupdatingtargetmodelafterthetrackofeachframeinrealtimethisalgithmis