基于人体特征的运动检测与跟踪-毕设论文.doc

PAGE\*MERGEFORMATIII基于人体特征的运动检测与跟踪摘要随着计算机视觉、模式识别等相关技术的发展，基于视频的运动人体检测和跟踪受到了多方关注，被广泛应用到银行、商场、停车场等对安全要求较高的地域监视中。基于视频的方法是唯一一种无入侵的运动捕获方法在海量视频数据检索、人机交互、视频监控、体育运动分析等领域有广泛的应用前景。实时场景下的人体跟踪会比较复杂，存在背景变化、遮挡、噪声及其他干扰因素，难以同时满足准确性、鲁棒性和实时性的要求。本文对单个摄像机且摄像机静止条件下的人体检测跟踪进行了研究，分析了现在已知的一些检测跟踪方法，采用背景减除法和投影比例检测法，提出了一种适用于室内室外环境，由运动物体检测、运动物体分析和人体运动跟踪3部分组成，使用单个静止数字摄像机得到德尔灰度图像序列作为输入，能够对单人或多人的运动行为进行监控的视频检测与跟踪系统。关键词：人体跟踪；人体检测；运动人体；detectionandtrackingofhumanaccordingtothefeatureofhumanABSTRACTWithcomputervision,patternrecognition,andotherrelatedtechnologies,basedonvideomotiondetectionandtrackingbythebodymuchattentionhasbeenwidelyappliedtobanks,shoppingmalls,parkinglotsandotherareasrequiringhighsecuritymonitored.Ontheonehandhumanmotioninvideomassivevideodataretrieval,human-computerinteraction,videosurveillance,sportsanalysisandotherfieldshaveawiderangeofapplications,andvideo-basedapproachistheonlynon-invasivemethodofmotioncapture.Itcontainssomeoftheotherproblems,suchashowtospecifythevideosequencetoidentifyaspecificbodymotion.Real-timehumantrackingscenarioswouldbemorecomplicated,thereisabackgroundchange,occlusion,noiseandotherinterferencefactorsdifficulttosimultaneouslysatisfytheaccuracy,robustnessandreal-timerequirements.Theproposedsystemisasuitableforindoorandoutdoorenvironmentsofvideodetectionandtrackingsystem,theentiresystemconsistsofamovingobjectdetection,motionanalysis,andhumanmotiontrackingobjectsinthreeparts.Thesystemisfullyimplementedinsoftwareapplicationthatrunsonanordinarypersonalcomputer,useadigitalcameratogetasinglestillimagesequenceasinputdelgray,youcanalsouseaninfraredimagesequences,capableofsingleormultiplayersportsbehaviormonitoring.Keywords:humantracking;humandetection;HumanMovement;目录摘要.....................................................................................................IABSRACT........................................................................................II第1章绪论......................................................................................11.1国内