基于MeanShift算法的视频球体目标的跟踪

2013毕业设计论文2013毕业设计论文编号：本科毕业设计（论文）题目：（中文）基于MeanShift算法的视频球体目标的跟踪（英文）MeanShift-basedTargetTrackingforMovingSphericalObjectinVideo分院信息科学与技术学院专业计算机与技术学院班级学号姓名指导教师职称完成日期2013毕业设计论文2013毕业设计论文诚信承诺我谨在此承诺：本人所写的毕业论文《基于MeanShift视频球体目标的跟踪》均系本人独立完成，没有抄袭行为，凡涉及其他作者的观点和材料，均作了注释，若有不实，后果由本人承担。承诺人（签名）：年月日2013毕业设计论文2013毕业设计论文摘要【摘要】在视觉跟踪领域，MeanShift算法以其实时性和鲁棒性强而著称，是一个非常优秀的算法，近几年来该算法发展迅速，有很大的发展前景。本论文主要研究基于核概率密度估计的MeanShift理论在视频序列图像中运动目标的识别与跟踪的应用。MeanShift算法避免了全局搜索，因此有很好的满足了实时性的要求。本文总结了核概率密度估计理论，并对MeanShift理论做了研究，对基于MeanShift理论的目标跟踪算法进行了深入研究，分析了其在视频跟踪中的优点，同时也指出了MeanShift算法的不足。其中，第二章详细介绍了MeanShift理论，第三章研究了MeanShift目标跟踪算法并在第四章对基于MeanShift算法的视频球体跟踪进行了试验与结果分析。【关键词】计算机视觉；目标跟踪；MeanShift；核函数；Bhattacharyya系数2013毕业设计论文2013毕业设计论文MeanShift-basedTargetTrackingforMovingSphericalObjectinVideoAbstract【ABSTRACT】Inthefieldofvisualtracking,MeanShift,whichisfamousforitsstrongreal-timeandrobustness,isaverygoodalgorithm.Thealgorithmdevelopedrapidlyinrecentyears,andhavegreatdevelopmentprospects.Inthispaper,ImainlyresearchtheuseofMeanShiftTheorybasedonkernelestimationinthemovingtarget.MeanShiftalgorithmavoidstheglobalsearch,soitcanmeettherequirementofreal-timeverywell.Thispapersummarizesthetheoryofkernelestimationofprobabilitydensity,andmakearesearchofMeanShiftTheory.ThenIanalyzeditsadvantagesintargettracking,andalsopointedoutitsdefectsatthesametime.Inthesecondchapter,thispaperintroducedtheMeanShiftTheoryindetail.InthethirdchapterIresearchedtheMeanShifttrackingalgorithmandmadeanexperimentofMeanShiftalgorithmandaresearchoftheresult.【KEYWORDS】ComputerVision;targettracking;MeanShift;Bhattacharyyacoefficient2013毕业设计论文2013毕业设计论文目录1绪论..............................................................................................................................................11.1引言..................................................................................................................................11.2国内外研究的现状............................................................................................................21.3目标跟踪综述...................................................................................................................21.3.1目标跟踪算法的分类...........................................................................................21.3.2特征的选择及提取..............................................................................................31.4论文内容安排...................................................................................................................32MeanShift理论.............................................................................................................................42.1引言..................................................................................................................................42.2MeanShift理论.................................................................................................................42.2.1MeanShift算法定义.............................................................................................42.2.2MeanShift基本思想.............................................................................................42.2.3扩展的MeanShift................................................................................................52.3MeanShift的物理意义.....................................................................................................72.4MeanShift算法的迭代步骤.............................................................................................72.5本章小结..........................................................................................................................83MeanShift目标跟踪算法............................................................................................................93.1目标图像的建模...............................................................................................................93.1.1初始帧目标特征模型的建立................................................................................93.1.2当前帧目标特征模型的建立................................................................................93.1.3特征匹配规则.....................................................................................................103.2MeanShift目标跟踪算法...............................................................................................113.2.1系数最大化........................................................................................................113.2.2算法描述............................................................................................................123.3本章小结........................................................................................................................124VisualC++2008编程实现Meanshift目标跟踪......................................................................134.1基于可视化面向对象的visualc++2008Express........................................................134.2OpenCV及其在目标跟踪中的应用..............................................................................134.3基于MeanShift的视频球体的跟踪的软件设计.........................................................134.3.1跟踪算法流