基于x射线成像和双目视觉技术的块冻虾虾含量测定分析

《基于x射线成像和双目视觉技术的块冻虾虾含量测定分析》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、江苏大学硕士学位论文ABSTRACTBlockfrozenshrimp(BFS)isanimportantseafoodinchinese，andthenetcontentofshrimp(NCS)isatlimportantqualityindexinBFS．mdetectionmethodofNCSisstilltraditionalnaturallythawing，drainning，andthenweighing．Tllisisnotonlytime．consuminganddamaging，butalsoC

2、annotbeappliedtotheindustrialautomationdetectionfastlyandnotconducivetotheimprovementofproductivity．Atthesametime，apraticalproblemaboutthe‘givingshortweight’ofNCSinBFSgetaincreasingnumberofcomplaintsbyconsumersrecently．Thus，arapid，non—invasivemethodinsolutionto

3、measuretheNCSinBFSWasproposedinthisarticle．Inthiswork，softX．ralyimagingandbinocularvisiontechnologywereemployedtonondestructivelyandrapidlymeasuretheNCSinBFS．Asfixingtheimagingparameters，thedifferencesofgraymeansofX-rayimagscouldbereflectedtheinsideinformationo

4、fBFSaccordingtotheattenuationruleofX·raypenetrationintothepotentialobjects．TheextemalinformationofBFSWasobtainedbybinocularvisiontechnologyonthebasisofmainlytheparallaxprinciple．ByfusingtheinternalandexternalinformationofBFS，theNCScanbeaccuratelycalculated．Them

5、aincontentsandconclusionsofthiSreasearchwerelistedasfollows：(1)nebestimagingparametersofX-raysystem，tubevoltageandtubecurrentofX-raygenerator,weredeterminedbycomparisionsofgrayhistogramsofX-rayimagesunderaseriesofdifferentimagingparametersforBFS；(2)Arelationshi

6、pbetweengraymeanandNCSWasinvestigatedbycomparingthedifferencesofgraymeanandX-rayimagesofBFSwiththeconditionofthesanlethickness．AndthisrelastionshipWasfeaturedwithhighlinearcorrelationandWasimpactedlittlebythefreezingtimebasedontheoptimalimagingparameters．Aserie

7、sofX．rayimagesofsamplesthatcharacteredwiththichnessof20m／n，25rnln，30InlTl，35InlTl，40mm，45mmanddifferentNCS，werecapturedrespectively．Thenagroupoflinearmodelsy=锻+6werecreatedbetweenaveragedgrey(x)andNCS(y)，withrelativecorrelationcoefficientsR20．9405，0．9874，0．9958

8、，0．9983，0．9897and0．9437，respectively．Thenthoselinearmodelswerevalidatedbyexternal20samples，whosethicknessandNCSweredifferfromonetotherest，withtherelativeerrorofpredictionlow