運動模糊交通標志檢測的研究.pdf

1、浙江大學(xué)碩士學(xué)位論文運動模糊交通標志檢測的研究姓名：高治雄申請學(xué)位級別：碩士專業(yè)：物理電子學(xué)指導(dǎo)教師：金韜20120314浙江大學(xué)碩士學(xué)位論文AbstractWiththedevelopmentofscienceandtechnology，intelligenttransportationsystem(ITS)hasbecomethesubjectofextensiveresearchathomeandabroadWeusecomput

2、ervisionsystemfortheacquisitionofenvironmentalinformationRoadsignisthemostimportanttrafficinformationSohowtoaccurateandidentifytheroadtrafficsignincomplexenvironmentbecomesthehotspotofITSandalsothemostdifficultproblemRoa

3、dtrafficsignimagesacquiredbyvehiclevisionsystemaredegeneratedbyunevenillumination，noise，andalsomotionblurduetovehiclemovementAlltheseproblemscausegreatdifficultiesinthedetectionandidentificationofthetargetsWetakeadeepstu

4、dyonthesubjectofmotionblurredimagerestorationAnimprovedalgorithmforrestoringnoisyandmotionblurredimageisproposedtopromoterobustnessandadaptabilityThealgorithmusescomplementarytemplatemethodofthemediantoestimatethenoiseen

5、ergyTheblurdirectionidentificationisbasedonimprovedRadontransform，withinthelogarithmicenergyspectrumdomainThenweidentifythefuzzylengthinthedoublespectralregionInordertofurtherimproveimageclarity，thehistogramequalizationm

6、ethodisusedtoimproveimagecontrastWeusetheregionsegmentationbasedonthedifferentcolorspacetodetectthetargetareaMethodsbasedontheRGBmodelandHSImodelarecomparedThesegmentationresultsarethenoptimizedusingthemorphologicalmetho

7、dforbetterresultsThedriverassistancesystemisstringentrequiredforreal—time，andatthesametimecan’toccupytoomuchspaceDSPbasedembeddedsystemsusingCMOSimagesensorastheimageacquisitionchipc鋤wellmeettheneedsItshighlevelofintegra