基于圖像處理的鋼板表面缺陷成像優(yōu)化與深度信息提取方法研究.pdf

1、武漢科技大學(xué)博士學(xué)位論文基于圖像處理的鋼板表面缺陷成像優(yōu)化與深度信息提取方法研究姓名：劉源泂申請學(xué)位級別：博士專業(yè)：機械設(shè)計及理論指導(dǎo)教師：孔建益20111218第Ⅱ頁武漢科技大學(xué)博士學(xué)位論文AbstractInthisdissertationaimingattheproblemsofimagingoptimizationandthedepthfeatureinformationextractioninthetechnologyofst

2、eelplatesurfacedefectsdetectionbasedonmachinevisionwimimprovingtheimagequalityasthestartingpoint’theinteractioneffectlawsandrelatedtheorieshavebeenresearchedwhichofthesteelplateproductionprocess，thesurfaceopticalproperti

3、esandtooptimumimagingofsurfacedefectsAndthekeytechniquesofsteelplatesurfacedepthextractionWgl“eachievedbasedonthemethodsofstereovisiontoenhancetheperformanceofthesteelsurfaceinspectionsystembasedonmachinevisionandprovide

4、atheoreticalbasis，andwidenedtheideasortechniquesofautomaticsteelplate’Ssurfaceinspectionsystem111emainworksandachievementsofthethesisaleasfollows：1)Buildinganaccuratesurfaceilluminationmodelformachinevisiontodescribethel

5、ightscatteringdistributionofsteelplatesurfaceThemodelisbasedonsteelproductionprocessandsurfaceopticalpropertiesandcombinedwith(BidirectionalReflectanceDistributionFunction)BRDFlightingmodeltheoryThroughtheexperimentsligh

6、tscatteringcharacteristicsofdifferentsteelplatesurface，therulesofrelationshipbetweenthelightincidenceanglesurfaceroughnessandlawsoflightscatteringunderaparticularlightsourceconditionswerefoundTheresultsshowedthatthereWas

7、allapparentspecularreflectionpeakonsteelsurfaceandsurfacelightscatteringWasinfluencedgreatlybylightincidenceangleandsurfaceroughness，ofthelawofexponentialdistributionfunctionsThustheimprovedsemiempiricallightscatteringma

8、thematicalmodelwhichbasedonroughnessfactorandsurfaceGaussiandistributionofmiotrplanecomponentshasbeenformed，andthroughnonlinearmodelfittingandoptimizationtodeterminetheparameters2)Onthebasisofanalysisandstudyonmanysample

9、sofsteelplatetypicalsurfacedefects，aimedattheoptimizationproblemofimagingopticalpatternforsteelplatesurfacedefects，wedesignedtheimagingexperimentschemeanddevelopmentthespecialexperimentequipmenttesearchthatquantifiesthee

10、ffectsofopticalimagingparameterstothesurfacedefectsimagesFurthercomprehensiveevaluationsystemofimagequalitybasedontheimages’multifeaturesWasproposedtoobtaintheoptimalimagingprogramforeverydefectFurthermore，throughtheopti

11、mizationimagingprocessmodelofsteelplatesurfacedefectseachsubimagingschemeswercanalyzed，andtheportfoliooptimizationimagingsystemdesignschemeWasobtainedAsthetheoriesresearchplatformofsteelsurfaceimagingfordefects，themodelp

12、rovidedabasistoimprovetheperformanceofthedetectionsystem3)InviewofachievingthreedimensionaldepthinformationextractionusingStereoVisionontheindustrialapplications，firstlyresearchedthet：atmeranonlinearcalibrationtechnology

13、BuildingthenonlinearcamerapinholemodelsforareascanCCDcameraandlineSCanCCDc鋤aabasedontheLENSdistortionmodel，determinedtheexteriorparametersandinteriorparametersofcameracalibrationUsingHALCONplanarcalibrationboardwhichWasc