基于視頻的運(yùn)動(dòng)物體檢測(cè)與提取方法研究.pdf

1、學(xué)校代號(hào)：學(xué)號(hào)：密級(jí)：10532S03102043普通湖南大學(xué)碩士學(xué)位論文y8926f】3基于視頻的運(yùn)動(dòng)物體檢測(cè)與提取方法研究掌僮宴重厶蛙當(dāng)；圭國(guó)塹墨9歪筵名廈驅(qū)建：鄒jE臻熬援墻差皇焦；讓箕扭當(dāng)通信堂隧童些芻整!鹽簋拯應(yīng)旦擅苤縫塞握童旦期；2QQ生且8目造塞答擻旦期!!Q堂生旦2Q旦簽蛙委雖金圭廑；奎』!堅(jiān)量三鎏耋蘭蘭墊望堡壘型皇堡堡童鎏塑薹ABSTRACTWiththedevelopmentofhightechnology，inte

2、lligentsurveillancesystemsareusedmorewidelyAsthecoretechnologyintheintelligencesurveillance，movingobjectsanalysisbasedonvideoconsistsofthemovingobjectsdetectionandretrieval，theobjectclassification，incidentdetection，actio

3、nidentificationandanalysis，andSOon，thedetectionandretrievalofmovingobjeetsisthefoundationandkeyofitInthedetectionandretrievalofthemovingobjects，thepresenceofshadowsinanimagecanleadtomisclassifytheobjectsormergethediffere

4、ntobjects，andbringthewrongresultsforthefollowingadvancedprocess，SOitcan’ttracktheobjectaccuratelyandgivethecorrectunderstandinganddescriptionoftheobjectsBasedonthesummaryandanalysisoftherelevantresearchworkshomeandabroad

5、，wemakeresearchonhowtodetectandretrievetheforegroundandhowtoeliminatetheshadowregioninthemovingobjectsdetectionandretrievalThemainresearchcontextsandresultsareasfollows：1Thebackgroundsubtractionmethodisusedtetrievethefor

6、eground，andthemixtureGaussianmodelisusedtomodelthebackground，duringthemodeling，theimprovedK—meansalgorithmsisappliedtoimprovethespeedofbackgroundmodeling；2Intheprocessofthebackgroundupdating，theupdatingalgorithmsbasedons

7、tatisticalaverageisusedtoupdatethebackground，comparedtothetraditionalbackgroundupdatingmethods，itimprovesthespeedofthebackgroundupdating；3TosolvetheproblemthatthedetectionmethodbasedonGaussianshadowmodeldetectstheshadowi

8、naccuratelyundersomesituations，anovelalgorithmbasedonthebodycolorvectormatchingispresentedtodetectandeliminatetheshadowFirst，aftertheforegroundisretrieved，theluminancetestisusedteduceparttheobjectregionthathashigherpixel

9、’SintensityintheforegroundthanthecorrespondingbackgroundSecondly，eachregion’Sorientationlineiscalculatedtoprejudgetheregionsthatcontaintheshadowregions，andtheregionsthatcontainshadowregionaremarkedThirdly，themarkedregion