基于高光譜成像技術(shù)的冷鮮羊肉品質(zhì)無損檢測研究.pdf

1、獨(dú)創(chuàng)_l生聲明本人聲明所呈交的論文是我個人在導(dǎo)師指導(dǎo)下進(jìn)行的研究工作及取得的研究成果。盡我所知，除了文中特別加以標(biāo)注和致謝的地方外，論文中不包含其他人已經(jīng)發(fā)表或撰寫過的研究成果，也不包含為獲得寧夏大學(xué)或其它教育機(jī)構(gòu)的學(xué)位或證書而使用過的材料。與我一同工作的同志對本研究所做的任何貢獻(xiàn)均已在論文中作了明確的說明并表示了謝意。研究生簽名：雖娩時間：汐f6年主月％曰關(guān)于論文使用授權(quán)的說明本人完全了解寧夏大學(xué)有關(guān)保留、使用學(xué)位論文的規(guī)定，即：學(xué)校

2、有權(quán)保留送交論文的復(fù)印件和磁盤，允許論文被查閱和借閱，可以采用影印、縮印或掃描等復(fù)制手段保存、匯編學(xué)位論文。同意寧夏大學(xué)可以用不同方式在不同媒體上發(fā)表、傳播學(xué)位論文的全部或部分內(nèi)容。(保密的學(xué)位論文在解密后應(yīng)遵守此協(xié)議)研究生簽名：五娘嘶帆秒，≤年至月巧日翩虢慢＼芝影帆勱名年多月Ⅻ日Nondestructivedetectionofchilledmuttonqualitybasedonhyperspectralimagingtechno

3、loAbstractInthisstudy，themuRonqualityparametersofTanHancrossbreedsheepintermsoftenderness，chromaticity，moisturecontent，storagetimeandthediscriminantofchilledandfrozenmeatwereevaluatedbyhyperspectralimagingtechnology(400—

4、1000nnl)combinedwithchemometricsmethod，whichcanprovidethetheoreticalandtechnicalsupportforonlinedetectionofanimalproductsThemainreseaRchcontentsandresultsareasfe『flows：(1)Comparedtotheoriginalspectrummodel，PLSRmodelsesta

5、blishedby400_1000nlnspectrumwithSavitzkyGolaysmoothingshowedbetterperformancetopredictthetendernessinchilledmuttonThecorrelationcoefficientofcalibrationandvalidationabouttendernesspredictionmodelsbasedon400100011111spect

6、rumwithSavitzky—GolaysmoothingwereO833and0865ThecorrelationCOe所cientofcalibrationandvalidationaboutcolorL4value(1uminance)colora木value(redness)andcolorb木value(yellowness)predictionmodelsbasedon400—1000nmspectrumwithSavit

7、zky—GolaysmoothingSNVandMSCwere0834，08M，0904andO736，0796，O898Comparedtotheoriginalspectrummodel，PLSRmodelsestablishedby400—1000nmspectrumwithpretreatmentapproachespresentedexcellentabilitytopredictthecolorparameterinchil

8、ledmutton(2)ComparedtoPCRmodelsoftheoriginalspectrum，PLSRmodelsbasedon400—1000nmspectrumhaveanexcellentabili夠topredictthemoisturecontentinchillednmtton，withwhichcorrelationcoefficientofcalibrationandvalidationwere0867and

9、0749，rootmeansquareerrorwas0707Comparedtotheoriginalspectrummodelandpretreatedspectrummodels，PLSRmodelsestablishedby4001000nmspectrumwithSavitzkyGolaysmoothingproducedthebesteffecttopredictthemoisturecontentinchfiledmutt

10、on，withwhichcorrelationcoefficientofcalibrationandvalidationwere0888and0784rootmeansquareerrorwas0696f3、TheLDAmodelwasestablishedforthewholebandspectrumof400—1000nlTl、)VhenthenumberofprincipalcomponentswasPC15，themodelwa

11、sbesttodeterminethecoldstoragetimeofmutton，andthetotaldiscriminantaccuracywas9813％Foramodelinthecaseofconstantprincipalcomponents，theeffectofmodelbasedontheoriginalspectrumwith6pointsPeaknormalizationwasbetterTotheLDAmod

12、elestablishedbasedonthe4001000nmspectrumwith6pointsPeaknormalization，thetotaldiscriminantaccuracyrateofthemodelwas9125％，whicheffectofmodelwasbetterthantheoriginalspectrumandotherpretreatedspectrummodelsComparedtothe、effe

13、ctoftheoriginalspectrummodelwithproperprincipalcomponentsofPC15andtheoriginalspectrummodelwith6pointsPeaknormalization，theresultsshowedthattheeffectoftheprincipalcomponentontheLDAdiscriminantmodelisbiggerthanthatofthespe