202光谱学与光谱分析 第35卷
[5] ZHANGWei,PANLei-qing,TUKang(张 伟,潘磊庆,屠 康).TransactionsoftheChineseSocietyofAgriculturalEngineering(农
业工程学报),2012,28(21):149.
[6] LuR,ArianaDP.PostharvestBiologyandTechnology,2013,81:44.[7] LiH,LiangY,XuQ,etal.AnalyticaChimicaActa,2009,648(1):77.
[8] AraújoMCU,SaldanhaTCB,Galv?oRKH,etal.ChemometricsandIntelligentLaboratorySystems,2001,57(2):65.[9] TIANHai-lei,LIHong-ru,XUBao-hua(田海雷,李洪儒,许葆华).ComputerEngineering(计算机工程),2013,39(4):222.[10] LIULu,WANGTai-yong(刘 路,王太勇).JournalofTianjinUniversity(天津大学学报),2011,44(9):803.
[11] WUJing-long,YANGShu-xia,LIUCheng-shui(吴景龙,杨淑霞,刘承水).JournalofCentralSouthUniversity?ScienceandTechnolo-[12] WANGXing-ling,LIZhan-bin(王兴玲,李占斌).PeriodicalofOceanUniversityofChina(中国海洋大学学报).2005,35(5):859.[13] LIXiao-lei,SHAOZhi-jiang,QIANJi-xin(李晓磊,邵之江,钱积新).SystemsEngineering—Theory&Practice(系统工程理论与实
践).2002,22(11):32.
[14] HUZhong-gong,LIJing(胡中功,李 静).ControlTheoryandApplications(控制理论与应用),2008,27(2):13.
[15] LIXiao-lei,LUFei,TIANGuo-hui,etal(李晓磊,路 飞,田国会,等).JournalofShangdongUniversity?EngineeringScience(山东
大学学报?工学版),2004,34(5):64.
gy(中南大学学报?自然科学版),2009,40(1):180.
Non-DestructiveDetectionResearchforHollowHeartofPotatoBasedonSemi-TransmissionHyperspectralImagingandSVM
HUANGTao,LIXiao-yu*,XUMeng-ling,JINRui,KUJing,XUSen-miao,WUZhen-zhongCollegeofEngineering,HuazhongAgriculturalUniversity,Wuhan 430070,China
Abstract Thequalityofpotatoisdirectlyrelatedtotheirediblevalueandindustrialvalue.Hollowheartofpotato,asaphysio-logicaldiseaseoccurredinsidethetuber,isdifficulttobedetected.Thispaperputforwardanon-destructivedetectionmethodbyusingsemi-transmissionhyperspectralimagingwithsupportvectormachine(SVM)todetecthollowheartofpotato.Comparedtoreflectionandtransmissionhyperspectralimage,semi-transmissionhyperspectralimagecangetclearerimagewhichcontainstheinternalqualityinformationofagriculturalproducts.Inthisstudy,224potatosamples(149normalsamplesand75hollowsamples)wereselectedastheresearchobject,andsemi-transmissionhyperspectralimageacquisitionsystemwasconstructedtoacquirethehyperspectralimages(390~1040nm)ofthepotatosamples,andthentheaveragespectrumofregionofinterestwereextractedforspectralcharacteristicsanalysis.Normalizewasusedtopreprocesstheoriginalspectrum,andpredictionmod-elweredevelopedbasedonSVMusingallwavebands,theaccuraterecognitionrateoftestsetisonly87.5%.Inordertosim-plifythemodelcompetitiveadaptivereweighedsamplingalgorithm(CARS)andsuccessiveprojectionalgorithm(SPA)wereuti-lizedtoselectimportantvariablesfromtheall520spectralvariablesand8variableswereselected(454,601,639,664,748,827,874and936nm).94.64%oftheaccuraterecognitionrateoftestsetwasobtainedbyusingthe8variablestodevelopSVMmodel.Parameteroptimizationalgorithms,includingartificialfishswarmalgorithm(AFSA),geneticalgorithm(GA)andgridsearchalgorithm,wereusedtooptimizetheSVMmodelparameters:penaltyparametercandkernelparameterg.Aftercompar-ativeanalysis,AFSA,anewbionicoptimizationalgorithmbasedontheforagingbehavioroffishswarm,wasprovedtogettheoptimalmodelparameter(c=10.6591,g=0.3497),andtherecognitionaccuracyof100%wereobtainedfortheAFSA-SVMmodel.Theresultsindicatethatcombiningthesemi-transmissionhyperspectralimagingtechnologywithCARS-SPAandAFSA-SVMcanaccuratelydetecthollowheartofpotato,andalsoprovidetechnicalsupportforrapidnon-destructivedetectingofhol-lowheartofpotato.
Keywords Hyperspectralimaging;SupportVectorMachine;Artificialfishswarmalgorithm;Hollowheart;Potato*Correspondingauthor
(ReceivedDec.25,2013;acceptedMar.15,2014)
半透射高光谱成像技术与支持向量机的马铃薯空心病无损检测研究
作者:作者单位:刊名:英文刊名:年,卷(期):
黄涛, 李小昱, 徐梦玲, 金瑞, 库静, 徐森淼, 武振中, HUANG Tao, LI Xiao-yu, XU Meng-ling, JIN Rui, KU Jing, XU Sen-miao, WU Zhen-zhong华中农业大学工学院,湖北 武汉,430070光谱学与光谱分析
Spectroscopy and Spectral Analysis2015(1)
引用本文格式:黄涛.李小昱.徐梦玲.金瑞.库静.徐森淼.武振中.HUANG Tao.LI Xiao-yu.XU Meng-ling.JIN Rui.KU Jing.XU Sen-miao.WU Zhen-zhong 半透射高光谱成像技术与支持向量机的马铃薯空心病无损检测研究[期刊论文]-光谱学与光谱分析 2015(1)
相关推荐: