资源描述:
《Metal surface defect detection using iterative thresholding technique 》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、nd2InternationalConferenceonCurrentTrendsinEngineeringandTechnology,ICCTET’14►561◄MetalSurfaceDefectDetectionusingIterativeThresholdingTechniqueM.SenthikumarDr.V.PalanisamyDr.J.JayaDepartmentofECE,AnnaUniversity,Principal,Principal,DindigulInfoInstituteofEngineering,AkshayacollegeofEngineering&Te
2、chnology,Coimbatore,Coimbatore,Tamilnadu.TamilnaduAbstract—Recently,surfacedefectsdetectioninmetalstheycomefromthesameimage.Tosolvethisproblemofplaysasignificantroleincomputervisionapplications.Anmultipleoutputsofagiveninputimagewithaspecificdefectefficientandaccuratedefectdetectionapproachistype
3、,avotingstrategyisproposedforthefinaldecisionimplementedinthispaper.Thedefectdetectiononmetalbasedontheone-versus-all[4]multipleclassifiers.surfaceisachievedbyiterativethresholdingtechniqueonmetalsurfaceimages.ThedefectregionsuchascrackandAnalgorithmfordetectingseamcracksinsteelplatesisshrinkageo
4、fthemetalsurfaceimageisdetectedbyillustratedin[5].ItusesGaborFilterandadaptivedoublebinarizationusingiterativethresholdingtechnique.Thethresholdmethodtodetectcracks.AnalgorithmstartswithexperimentalresultsarecarriedoutbyusingrealtimemetalthesegmentationofacquiredimageandthenGaborfunctionsurfaceim
5、agesandsatisfactoryperformanceisachievedbytheisappliedforcreationofGaborfilterbank.Thenadaptiveproposeddefectdetectiontechnique.doublethresholdingprocessisusedtoobtainabinarizedimage.Keywords:surfacedefectdetection,iterativethresholding,binarization,metalsurface.Anautomaticdefectdetectiontechniqu
6、eforceramictilesisdescribedin[6].TheimageprocessingoperationssuchasI.INTRODUCTIONimageacquisition,imageenhancement,noisereductionandedgedetectionareused.Afterthedetectionofdefects,Thetechniquesusedforsurfacedefectdetectionarevariousalgorithmsareusedtoclassifythedefects.Theclassifiedintofourdiffer
7、entapproaches.Theseapproachesreflectionpropertiesacrosstheflawlesssurfacechangesarestatistical,modelbased,structuralandfilterbased.Instronglyduetoscaleonthesurface.Sinceintensityimagingstatisticalapproachhistogramprope