图像去雾算法研究及其在交通监控视频中的应用-控制科学与工程专业论文.docx

AbstractIn foggyconditions，theimagescollectedby theoutdoorcomputervisionsystemwillbegreatlydegraded，suchascontrastdecline，colordistortion．Inrecentyears，thefogweatherfrequentlyappears，thisresultinasignificantnegativeimpactontheuseofoutdoorcomputervisionsystem，SOitsofgreatvaluetohaveastudyonimagedefogging．Basedonfogimageenhancementandfogimagerestoration,paperhasastudyonimagedefoggingtechnology,establishestheguidelinesthat‘‘ensurethedefoggingeffectandenhancethespeedasfaraspossible”，makessomeimprovementsfortherelatedalgorithms．Applyingtheimagedefoggingalgorithminpracticalapplicationtoverifiesit’Seffectiveness．Themainworkofthepaperisasfollows：(1)Intermsofimageenhancement，paperanalyzesthedefoggingeffectofhistogramequalizationalgorithmandRetinexalgorithm．Enhanced byRetinexalgorithm，thefoggyimagewillappearcolordistortion,aimingatthisdefect，paperusemulti．scaleRetinexwhichbasedonHSVcolorspacebutnotRGBcolorspacetoenhancethefoggyimage．InHSVcolorspace，lightnesscomponentVandhuecomponentHareuncorrelated，basedonthischaracteristics，whenenhancingthecomponentVcomponentHremainsunchanged．Experimentalresultsshowthattheproposedalgorithmhaveabetterenhanceresukoncolorfulimage，atthesametime，asonlytoenhancecomponent钐theproposedalgorithmisfaster．(2)Inrespectofimagerestoration,thepaperisfocusingonthestudyingoftheimagedefoggingalgorithmbasedondarkchannelprior．Inordertoeliminatethehaloeffect，originalalgorithmusethesoftmattingmethodtorepairthetransmittancemap，thisleadtothecomplexityofthealgorithmhighly，andcomputingspeedslow．Inthispaper，smoothingthetransmittancemapisprocessedbyadaptivebilateralfilteringalgorithmwhichreplacetheoriginalsoftmatting，andusethedown-samplingmethodtomakethespeedfaster．SimulationresultsvalidatethattheproposedalgorithmCaneffectivelyeliminatethehaloeffectandthespeedofthealgorithmhasbeengreatlyimproved．Ⅱ万方数据(3)Basedonthetheoreticresearch，thedefoggingalgorithmisappliedto thesituationofroadtrafficmonitoring．Basedonthefeaturesoftrafficvideomonitoring，usingasuitabledefogalgorithmi