鋁箔紙包裝鴨肫異物檢測關(guān)鍵技術(shù)研究.pdf

1、南京林業(yè)大學(xué)碩士學(xué)位論文鋁箔紙包裝鴨肫異物檢測關(guān)鍵技術(shù)研究姓名：洪冠申請學(xué)位級別：碩士專業(yè)：檢測技術(shù)與自動化裝置指導(dǎo)教師：趙茂程20070601AbstractFoodforeignbodiesharmthehumanhealthandholdbackthefoodindustrydevelopmentTodetectfuodbycomputervisionandimageprocessingtechnology，theexaminat

2、ionhastheadvantagesofnondestructive，fastaccurate，andreliableereTakingaluminumfoilpackingduckliverasobjectthisarticlemakesresearchontheradiographicimageofduckliverimagebyXrayinordertoachievetherecognitionandextractionfrom

3、duckliverofforeignbodiesbasedonthecomputervisionandtheimageprocessingtechnologyⅢSarticlehasstudiedXraydigitalimageradiographicoptimizationandemendationmethodAndthecharacteroftheradiographicimageofduckliveriSanalyzedXrayc

4、oefficientofattenuationpictureelementvaluemathematicalmodeliSdevelopedtoestimateblankimageintensity、duckliverandforeignbodiesatdifferentvoltagesandcurrentsnlemethodofvaltleishigherthanO995；Inordertoguaranteethegoodeffect

5、oftheforeignbodiesSegmentandtoenhancetheforeignbodiesinspectionspeed，amethodofsettingXrayparameter(tubevoltage，tubeelectriccurrent)isproposedinthispapern圮characteristiesofthenoisefromImageIntensifierandtheCCDcameraoftheX

6、rayradiographicimagingsystemarcanalyzedInordertoimprovetheimageacquisitionqualityandreducethenoisedisturbing，multiframemeansfilterprocessing，reducingintheimagethestochasticnoise，iSproposedInordertoensuretheforeignbodieso

7、fduckliverimagespecificclearly，5x5GaussfiltermethodisemployedintheduckliverimagefurdenoisingTlIiSthesishasstudiedintheforeignbodiesofX—rayradiographicduckliverimageautomaticiudgmentmethodtoseparatetheduckliverfromtheback

8、groundbyusingOTSUmethod；Underthelimitationtothesizescopeandthegreatgradientchangestoforeignbodiesofduckliverimaget11eforeignbodiesalesuccessfullyseparatedfromduckIiverbyActiveContourModelmethodTlliSapproachisofsuchhighin

9、spectionprecisionthatitcallaccuratelyascertainforeignbodiessuchasstone，boneandmetalwireabove02mmindiameterOnthebasisofresearches，throughtheshapeandimagegrayscaleofduckliver’Sforeignbodies’characteristics，alleigenvectorth

10、atfactuallyreflectedforeignbodies’essentialpropertiesandtheircomputationiSpresentedOnsolutionofforeignbodies’classification，theselforganizedandselfadaptivebackpropagation①P)neuralnetworkalgorithmisusedtointelligentlyreco

11、gnizetheimageofduckliver’Sforeignbodies11坨above—mentionedsolutiontoforeignbodies’recognitionofduckliverinXrayradiographicimagepassedsimulationandamulation晰thLimitedsamplesimulationandtheresulthasobtainedaperfecteffc“111e

12、XrayautomaticinspectionsoftwaresystemfurduckIiver’SforeignbodiesisdevelopedundertheVisual科斗netplatformnliSsoRwaresystemcanelementarilyachievetherecognitionOftlleimageofduckliverinXrayandfulfillthebasicfunctionofthemaagep