基于混凝土無損探傷的超聲CT圖像分析和處理.pdf

1、武漢理工大學碩士學位論文基于混凝土無損探傷的超聲CT圖像分析和處理姓名：肖敏芳申請學位級別：碩士專業(yè)：應(yīng)用數(shù)學指導(dǎo)教師：李卓球20060501AbstractTheresearchinthispaperissupportedbytheNSFC’sFunding：SmartConcreteanditsStructure(ProgramNo：50238040)。Basedontheconcretenon—destructivetesting

2、technologytheprinciplesandexperimentalmethodsofultrasoundCTareexplored，therebylaunchingaseriesofimageanalysisandpost—processingThemainconclusionsare“1Discussiononcharacteristicsofconcrete131trasonicCTimages：Toacertainext

3、ent，cTimagescanreflecttheexistenceandlocationofintemalflaws。butcannotaccuratelyrepresentthesizeandshapeWhilecrimagestothesizechangearemoresensitive，butitisverydifficulttoidentifytheshapeofconcreteflaws2CTimagequaUtyanaly

4、sisoftheeffectsofnumerousfactors：Forexample，thecombinationsofflaws，therelationshipbetweenthenumberofsurveypointsandthenumberofimagingmodules，inversemodel，calculationaccuracyAndgiventhecorrespondingsolutions，asfaraspossib

5、leeliminatetheCTimagesof”pseudolike”3Establishingatask—basedCTimagequalityevaluationsystem：Fromthe。resolvingpowerreadabilityandreliabi娃tyoftheCTimage，thequalityofCTimageisdiscussed，thengivingasemiquantitativecriterion4Re

6、spectivelyweusemedianfilterclusteringanalysisandvariancecutofftechniquestoprocesstheimages，andcomparethre》emethodsIntegratedtechnology奴theexperimentalresultsdealingwith，andenhancethereliabilityoftheimagesmakingflawsregio

7、nandbackgroundsregiondifferencesapparent5WeusetraditionalmethodstodivideCimagesfindingitdifficultfortheborderdeterminationoftheconcreteflawsImagepartitiontechnologybasedonthebordertransitionzoneisintroduced，themethodsuse