异纤在线检测与清除关键技术的分析

《异纤在线检测与清除关键技术的分析》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、AbstractAppearancedefects，colorshadingandotherproblemsoftenariseduringthetextileprocessofforeignfibers，whicheffectthegradingofcoRon，spinningandfabricqualityseriously．Currentlyforeignfiberdetectionismostlydetectedbehindthecottonopening，whenthefibershavebeenpaper,thepr

2、ocessofforeignfiberstestingscattered．Soit’Sdifficulttodetect．Intheispositionedbeforetheopeningprocess·Themajorcontentsasfollowing：Basingonthestudyofforeignfibersdetectionstatushomeandabroad，theprocessandtechnicalrequirementsbeforetheopeningofforeignfibersisproposed·T

3、hen．theoverallstructureofthedetectionsystemisdesigned，whichincludesthestructuraldesignandcontrollerdesignofbreakingup，imageacquisitionandprocessing，andtherejectingofforeignfibers．Finallytheprototypewasproduced，whichcandetectthesmallestsizesheet6mm2andlinear1mm．Curren

4、tly,therunningoftheprototypeisverywell．Forexample：thinmaterialsuchashairCanreachmorethan85％．Accordingtotheperformancerequirementsofforeignfiber，theimagecollectionsystemwithproperpixelandresolutionwassetup’andtheadjustablestructuralofcameraandlightsourceWasdesigned，wh

5、ichsatisfiedtherequirementofcameralensfixedandadjustmentrequirements·ThekeytechnologyofforeignfiberimageprocessingsystemistofindingthedifferencesbetweenforeignfibersandcoRonfibers．Throughthecomparingandanalysisofthealgorithmsinvariousstages，properimageprocessingalgor

6、ithmsisin仃oduced·Aftergray。levelprocessmg,Medianfilteringmethodisusedtodenoiseimage，andtheidentificationofforeignfibersisachievedthroughtheedgedetection．Theextractingobjectboundaryinformationachievestheobjectivespositioningoftheforeignfibersincottonlayer．Accordingtoo

7、wncharacteristicsofthistopic，thecontrolsystemwhosecoreISPLCisusetocontrolrawcottonforeignfiberdetectionandcontrolsystemISdesignedinthepaper．Thepositioningsystemofforeignfibersinrawcottonpneumaticdetectionisanalyzedtoachievetimelyandaccuratedetectionoftherawcottonfore

8、ignfibersonline．Experimentsshowthattheforeignfibersdetectingsystemwithabovetechnologieshashigherdetectingability,andcandetectbefore