Real-time tracking of deformable objects based on MOK algorithm

在线阅读 下载PDF 导出详情
摘要 ThetraditionalorientedFASTandrotatedBRIEF(ORB)algorithmhasproblemsofinstabilityandrepetitionofkeypointsanditdoesnotpossessscaleinvariance.Inordertodealwiththesedrawbacks,amodifiedORB(MORB)algorithmisproposed.Inordertoimprovetheprecisionofmatchingandtracking,thispaperputsforwardanMOKalgorithmthatfusesMORBandKanade-Lucas-Tomasi(KLT).ByusingKalman,theobject’sstateinthenextframeispredictedinordertoreducethesizeofsearchwindowandimprovethereal-timeperformanceofobjecttracking.TheexperimentalresultsshowthattheMOKalgorithmcanaccuratelytrackobjectswithdeformationorwithbackgroundclutters,exhibitinghigherrobustnessandaccuracyondiversedatasets.Also,theMOKalgorithmhasagoodreal-timeperformancewiththeaverageframeratereaching90.8fps.
机构地区 不详
出版日期 2016年02月12日(中国期刊网平台首次上网日期,不代表论文的发表时间)
  • 相关文献