基于圖像處理的瀝青混合料分割與統(tǒng)計(jì)研究.pdf

1、西南交通大學(xué)碩士學(xué)位論文基于圖像處理的瀝青混合料分割與統(tǒng)計(jì)研究姓名：謝池鑫申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：馬永強(qiáng)201206西南交通大學(xué)碩士研究生學(xué)位論文第1I頁(yè)AbstractTheasphaltpavementisthemainpavementofseniorandtimesseniorroadinChinatheasphaltmixtureisamajorcomponentoftheasphaltpavement，

2、itsqualitydirectlyaffecttheservicelifeoftheroadAsphaltnlixtllretestingandstatisticaltechniqueshaveimportantsignificanceintheresearchofasphaltpavementTheasphaltmixtI】rebasedontheCTscanhavelessnoise，highdefinition，cleargra

3、inboundaries，itisrelativelyeasytomanaged，howevertheneedforspecializedequipment，thecostisrelativelyhigh，dimensionalasphalt血xtIlrebasedonthecamera，therearemanyunfavorablefactors，segmentationandstatisticalresultsarenotsatis

4、factorycannotbeeffectivelyseparatedasphaltmixtl】rethereisaconnectionparticlesThisthesisdetectmorecomplexcameraasphaltmixhlre，segmentparticlesintheimage，thecorrespondingstatisticalresultsgivenbythesegmentationresultsFirst

5、，thisthesisanalyzesamorecomplexprofileimageoftheasphaltmixtureshotbydigitalcamera，duetomanydisadvantagesoftheoriginalimage，thusimagepreprocessingisused，thenGaussiannoiseisremovedfromtheimagebytransformingitintothegraysca

6、le，removingtheunevenilluminationduetotheexternalenvironment，sharpeningtheimagefuzzyedgeinformation，segmentingthegrayimagebasedonOTSUthresholdsegmentationmethodSecond，removingthewavenoising，detectingtheedgeoftheasphaltIIl

7、ixtIlreimagebasedonthresholdsegmentationmethodcombiningthetwobinaryimagesAnalyzingthebinaryimageexistedconnection，F(xiàn)irstlyinordertomaintaintheshapeoftheaggregates，theimageiscardedoutsmallscalemorphologicaloperationsfordis

8、connectingthesmallconnection，thenproposingaiterativesegmentationalgorithmbasedonrelationshipbetweentheconcavepointandtheedgetemovethebigconnectionFirstlyaccordingtothebigconnectionisalwaysconcavecharacteristiCS，usingther

9、elationshipbetweentheconcavepointandtheedgetodoiterativesegmentationAnalyzingthesegmentedbinaryimageoftheasphaltmixture，F(xiàn)irstdeterminethestatisticalmethodsoftheaggregateparticles，onthisbasis，givetheoverallgeometricfeal：u

10、rcsandshapefeaturesoftheimage，drawthegradingcBrveofasphaltmixtureTheexperimentalresultsshowthattheasphaltmixtureinitialtwosegmentationCanbeabinaryimageoftheglueregionalandaggregateparticles，theseparationofconnectedpartic

11、lesabletolargelymaintainthenumberofsmallparticlesandlargeparticleswillnothavetoomuchdeformation，onlyasmallportionofparticlestherewillbeoversegmentationandundersegmentationphenomenonThestatisticalresultsCanaccuratelycalcu