资源描述:
《Development of Real-Time Vision-Based fabric inspection system》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、IEEETRANSACTIONSONINDUSTRIALELECTRONICS,VOL.52,NO.4,AUGUST20051073DevelopmentofReal-TimeVision-BasedFabricInspectionSystemChe-SeungCho,Byeong-MookChung,Member,IEEE,andMoo-JinParkAbstract—Qualityinspectionoftextilefabricproductsisanim-fabricdefectsiscr
2、iticallyimportantinthetextileindustry.Inportantproblemforfabricmanufacturers.Thispaperpresentsanparticular,ifthereisadefect,itreducesthepriceofthefabricautomaticvision-basedsystemforthequalitycontrolofwebtex-by45%–65%.Toincreasetheoverallquality,homog
3、eneityoftilefabrics.Typicalwebmaterialis13mwideandisdrivenwithfabric,andreliability,anautomatedvisualinspectionsystemisspeedsrangingfrom20to200m/min.Atpresent,thequalityin-spectionprocessismanuallyperformedbyexperts.However,theyneededforbetterproducti
4、vity.Atypicalwebmaterialis1–3mcannotdetectmorethan60%oftheoveralldefectsforthefabricwideandisdrivenwithspeedsrangingfrom20to200m/min.ifitismovingfasterthan30m/min.ToincreasethequalityandEveninthebestcases,expertscandetectnomorethan60%homogeneityoffabr
5、ics,anautomatedvisualinspectionsystemisoftherealdefects,andtheycannotdealwithfabricwiderthanneededforbetterproductivity.Currently,theexistinginspection2mandmovingfasterthan30m/min.Manyattemptshavesystemsaretooexpensiveforsmallcompanies.Inthispaper,aPC
6、-basedreal-timeinspectionsystemisproposedwithbenefitsofbeenmadetosolvetheseproblems.Forexample,Cohenetlowcostandhighdetectionrate.Theproposedalgorithmshowedal.[6]usedMarkovRandomField(MRF)modelsfordefectgoodresultsforseveraltypesoffabricdefects.inspect
7、ionoffabricsurfaces.ChenandJain[7]usedastructuralIndexTerms—Automatedinspection,fabricdefects,textileapproachtodetectdefectsintexturedimages.Dewaeleetal.fabricinspection,vision-basedsystem.[8]usedsignalprocessingmethodstodetectpointdefectsandlinedefec
8、tsintextureimages.Atalay[9]hasimplementedanMRF-basedmethodonaTMS320C40parallel-processingI.INTRODUCTIONsystemforreal-timedefectinspectionoffabrics.ThemajorityECENTLY,opticaldevicesarebeingincorporatedintoofautomatedinspectionsystemscurrentlyav