运动目标检测与跟踪及其在视频监控系统中的应用.pdf

Abstract Asthevideosurveillancewiththeconstant mereuse system expansion，the all ofhumantoachieve on—line is monitoringpoints monitoringalready for surveillance video many impossible surveillancewithautomaticorsemi—automaticfunctionsis system monitoring thefuture trendsofthevideosurveillance development system． onthe Thisthesisfocuses intelligentmonitoringsystemmovingtarget the on detection，tracking，andpracticalapplicationmonitoringsystem． Thisthesis thecurrent detection analyzes typicalmovingtarget algorithm． Forthecurrent to monitoringsystemneeds，appropriateimproveexisting the oftheframedifferencemethodand combining algorithms．By adaptability of difference the detectionof theaccurate backgroundmethod，achieverapid movingtargets． Thisthesis and the analysiscomparestypicalmovingtargettracking thenfocuseson basedonMeanShiflandCamShifl algorithms，and algorithms thisthesisconductedaseriesof on MovingTargetTracking．Then improvements the ofthe forthe themethod shortcomingsalgorithmCamShift．First，improved ofCamShiflback calculationacombinationofscene projection through a differencebetweenand