视频监控中运动人体检测与异常行为分析分析-analysis of moving human detection and abnormal behavior in video surveillance.docxVIP

AbstractIntelligentvideosurveillanceintroducesthetechnologyofcomputervisionandpatternrecognitiontechnologyinthevideosurveillance,whichistakingmoreandmoreattention.Theanalysisofhumanabnormalbehaviorinvideoisanincreasinglyimportantresearchdirectionofintelligentvideosurveillancesystem.Itcanachieveintelligentdetectionofhumanabnormalbehaviorbyusingthetechnologyofcomputervision,completesecuritytasksefficiently,andsavealotofhumanandmaterialresources.Thearticleismainlyformovinghumanabnormalbehavioranalysisinvideo,whichisspecificallyrelatedtomovinghumanbodydetection,trackingandabnormalbehaviordetection.Inthemovingobjectsdetection,inordertoextracttheforegroundtarget,thepaperproposesbinarymaskbackgroundmodelmethod..Bycontinuouslyupdatingthebackgroundtemplatewhenthetotalof0isinasmallvalue,theinitialbackgroundisestablished.Then,usingbackgroundsubtractionmethodextractsmovingobjects.Atthesametime,thebackgroundisupdatedadaptivelyduringobjectsdetectioninordertoreducetheimpactofexternalfactors.Inobjectstracking,thepapergivesatrackingmethodofadjacentframesareamatchandgrayscalematch.Inthebinaryadjacentframes,movingtargetownsalotofoverlapareatoachieveobjectsmatchingandtracking.Ifobjectsmergeorsplit,usingKamanfilteringandgrayscalehistogrammatchingmethodtracksobjects.Intheanalysisofabnormalbehavior,thepaperanalyzesbyextractinghumanfeatures(humancenterandlength-widthration)andhuman-moving-trajectorytothedetectionofabnormalbehavior.Therectanglelength-widthratioistakenasthecriteriaofjudgingthebodyfall.Thehoveringbehaviorisdetectedbasedonthedirectionanglechangeofmovementtrajectory.Intheremnantsdetection,thesystemanalysestherelationofblobs,andgetsthemovingstateofrelatedblobs,accordinglyjudgesthattheblobbelongstoremnants.ThepaperusesVisualC++6.0developmenttoolsandOpenCVcomputervisionlibrarytoverifytheperformanceofthealgorithm.Theexperimentalresultsshowthatthedetectionalgorithmcanquickly,accuratelyachievetheidentificationofhumanmovementabnormalbehaviorinthevideo.Keywords:VideoSurveillance,Abnorm