预览加载中，请您耐心等待几秒...

一种基于隐马尔可夫模型的目标轨迹跟踪算法.docx

预览
1/4
2/4
3/4
4/4

在线预览结束，喜欢就下载吧，查找使用更方便

下载提示 文本预览

如果您无法下载资料，请参考说明：

1、部分资料下载需要金币，请确保您的账户上有足够的金币

2、已购买过的文档，再次下载不重复扣费

3、资料包下载后请先用软件解压，在使用对应软件打开

一种基于隐马尔可夫模型的目标轨迹跟踪算法 Abstract Inrecentyears,withtherapiddevelopmentofcomputervisionandmachinelearning,theresearchonobjecttrackingalgorithmhasbeengreatlyimproved.Inthispaper,weproposeatargettrajectorytrackingalgorithmbasedonhiddenMarkovmodel.Firstly,weintroducethebasicprincipleandmathematicalmodelofhiddenMarkovmodel.Then,weanalyzetheproblemsexistingintraditionalobjecttrackingalgorithmsandthereasonsfordifficulttrackingofcomplexdynamictargets.Finally,ourproposedalgorithmispresentedindetail,includingtargetrepresentation,targetstateestimation,andtargettrajectoryprediction.Experimentalresultsshowthattheproposedalgorithmcaneffectivelytrackthetargettrajectorywithhighaccuracy. Keywords:ObjectTracking,HiddenMarkovModel,TargetTrajectory,StateEstimation Introduction Objecttrackingisoneofthemostimportanttopicsincomputervisionandithaswideapplicationsinmanyfieldssuchasvideosurveillance,autonomousdriving,androbotics.Traditionalobjecttrackingmethodsmainlyrelyonhand-designedfeaturesandoftenfailwhenfacedwithocclusion,illuminationchanges,andcomplexscenes.Withthedevelopmentofdeeplearning,someremarkableprogresshasbeenmadeinobjecttracking,butitstillfacesmanychallenges. ThehiddenMarkovmodel(HMM)isapowerfultoolformodelingtimeseriesdataandhasbeenwidelyusedinthefieldofspeechrecognition,bioinformatics,andnaturallanguageprocessing.Inrecentyears,theHMM-basedobjecttrackingalgorithmhasbeengraduallyappliedinthefieldofobjecttrackingduetoitsabilitytodealwithcomplexdynamictargets. ThepurposeofthispaperistoproposeanobjecttrackingalgorithmbasedonthehiddenMarkovmodeltoimprovetheaccuracyandrobustnessofobjecttracking.Therestofthispaperisorganizedasfollows.Section2introducesthebasicprinciplesandmathematicalmodelsofthehiddenMarkovmodel.Section3analyzestheproblemsexistingintraditionalobjecttrackingalgorithmsandthereasonsfordifficulttrackingofcomplexdynamictargets.Section4proposesanobjecttrackingalgorithmbasedonthehiddenMarkovmodel.Section5presentsexperimentalresultsandanalysis.Finally,Section6concludesthispap