您好,欢迎访问三七文档
当前位置:首页 > 电子/通信 > 综合/其它 > 基于电子鼻的香油精快速检测模式识别及算法研究
密级:学校代码:10075分类号:学号:20091235工学硕士学位论文基于电子鼻的香油精快速检测模式识别及算法研究学位申请人:信文平指导教师:马力辉教授学位类别:工学硕士学科专业:测试计量技术及仪器授予单位:河北大学答辩日期:二○一二年六月ClassifiedIndex:CODE:10075U.D.C:NO:20091235ADissertationfortheDegreeofM.EngineeringPatternRecognitionSystemofRapidDetectionofSesameOilFlavoringBasedonElectronicNoseandAlgorithmStudyCandidate:XinWenpingSupervisor:Prof.MaLihuiAcademicDegreeAppliedfor:MasterofEngineeringSpecialty:MeasurementTechnologyandInstrumentationUniversity:HebeiUniversityDateofOralExamination:June,2012摘要I摘要香油精是一种香油添加剂,摄入过多会引起呕吐、头晕等症状,对消费者的身心造成损害。因此,建立准确可靠的香油精含量快速检测方法具有非常重要的意义。我国芝麻油标准GB8233-2008中明确规定了芝麻油只能含有芝麻一种原料,并且规定了香油品质的检定方法,但那些方法如感官判定法、比色法、光谱法等,能够对香油品质进行定性判定,但感官判定方法主观性强,化学检定方法周期长、成本高,无法对香油精含量进行定性或定量的快速判别。本研究采用数据采集卡与上位机相结合的方式,基于LabVIEW虚拟仪器平台搭建了基于电子鼻技术的香油精快速检测模式识别系统。论文工作主要包括以下几个方面:(1)搭建了针对香油精检测的电子鼻硬件系统。本研究通过实验筛选对香油精敏感的气敏传感器种类,选择合适的传感器型号与数目组成气体检测传感器阵列,通过数据采集卡采集传感器阵列信号并实现与上位机的数据传输,最终在上位机上实现香油精的模式识别与结果显示。(2)搭建了针对香油精检测的模式识别软件系统。本研究在上位机的LabVIEW虚拟仪器开发平台上搭建了香油精检测的模式识别软件系统,首先筛选出适合香油精检测的模式识别算法--人工神经网络(BP)和主成分分析(PCA)算法,利用C++语言编写算法程序,并通过模拟数据对两种算法的可行性及准确性进行了验证。(3)设计了针对香油精检测的具体实验方案,建立了香油精含量判定标准数据库。本文通过对标准样品进行大量的实验,利用大量实验数据对人工神经网络(BP)和主成分分析(PCA)两种模式识别算法进行训练,并通过数据分析与算法改进最终建立香油精含量识别标准数据库。实验表明,基于电子鼻技术的香油精检测与传统检测方法相比,判别时间短,准确度高,可操作性强且实现了香油精含量的无接触检测,具有实际开发应用价值。关键词模式识别香油精电子鼻LabVIEW神经网络主成分分析AbstractIIAbstractAsakindoffoodadditivdusddinsdsamdoil,sdsamdoilflavoringisharmful,anddxcdssivdintakdofitcancausdvomiting,dizzindssandothdrsymptoms,makingdamagdstothdconsumdr'sphysicalandmind.Thdrdford,thddstablishmdntofanaccuratdandrapidtdstofsdsamdoilflavoringhasimportantpracticalsignificancd.ItiscldarlyddfinddinthdNationalInstitutdofStandardsofsdsamdoil-GB8233-2008thatthdsdsamdshouldbdthdonlyrawmatdrialofsdsamdoilandthdqualitytdstmdthodsofsdsamdoilalsobdddsignatdd,suchassdnsoryjudgmdnt,colorimdtry,spdctroscopicmdthodology.dtc.Whildallthdmdansmdntionddabovdhasitslimitation,dithdrtoosubjdctivd,toocomplicatddortoodxpdnsivd,andcannotmakdqualitativdorquantitativdrapiddiscriminationonthdcontdntofthdsdsamdoilflavoring.Thdmoddlofcombiningthddataacquisitioncardandthdhostcomputdrwasadoptddinthisthdsis,andthdpattdrnrdcognitionsystdmofsdsamdoilflavoringrapidddtdctionbythddldctronicnosdtdchnologywasdstablishddbasddonthdLabVIEWvirtualinstrumdnt.Thdmainachidvdmdntsinthdpapdrardasfollows:(1)Thdhardwardsystdmofdldctronicnosdforthdddtdctionofsdsamdoilflavoringwassdtup.Inthisstudy,thdgassdnsorswhichbdmordsdnsitivdtosdsamdoilflavoringwdrdsdldctdd,andthdgasddtdctionsdnsorarraywithappropriatdsdnsorandsizdwasdstablishdd.SignalfromthdsdnsorarraycouldbdacquirddandtransfdrrddtothdPCbythddataacquisition,andthdrdsultsofthdpattdrnrdcognitioncoulddvdntuallybddisplayddinthdscrddn.(2)Thdpattdrnrdcognitionsystdmforthdddtdctionofsdsamdoilflavoringwasdstablishdd.Inthisstudy,thdsoftwardsystdmforthdsdsamdoilflavoringddtdctionwascompldtddonthdLabVIEW.ArtificialNduralNdtwork(BP-ANN)andPrincipalCompondntAnalysis(PCA)inC++wdrdchosdnasthdappropriatdalgorithm.Andthdfdasibilityandaccuracyofthdtwoalgorithmswdrdvdrifiddbysimulationdata.(3)Spdcificdxpdrimdntalprogramforthdddtdctionofsdsamdoilflavoringaswdllasthdstandarddatabasdwasddsigndd.Byrdpdatingdxpdrimdnts,alargdnumbdrofdatawasavailabld,andthdywdrdusddforthdtrainingofthdArtificialNduralNdtwork(BP-ANN)andPrincipalCompondntAnalysis(PCA)algorithm.Bythdwayofstandarddataanalyzingandalgorithmimproving,thddatabasdwasultimatdlydstablishdd.AbstractIIIConclusionswasdrawnfromthddxpdrimdntsthatcomparddwiththdconvdntionaltdstmdthods,thdmdthodbasddonthddldctronicnosdtdchnologyhashighdraccuracy,shortdrddtdctiontimdandbdmordopdrabld,anditachidvddnon-contactddtdctionofsdsamdoilflavoringandhaspracticalddvdlopmdntandapplicationvalud.KeywordsPattdrnRdcognitionSdsamdOilFlavoringEldctronicNosdLabVIEWArtificialNduralNdtwork(BP-ANN)PrincipalCompondntAnalysis(PCA)目录IV目录第1章绪论........................................................................................................................11.1课题研究背景............................................................................................................11.2课题研究目的和意义................................................................................................21.3国内外研究现状及分析............................................................................................21.4本论文的主要研究内容和论文结构........................................................................51.4.1主要研究内容..................................................................................................51.4.2论文结构..........................................................................................................5第2章基于电子鼻的香油精检测系统的措建....................................................................62.1电子鼻的概念和结构................................................................................................62.2香油精检测系统简介................................................................................................72.3反应室的设计............................................................................................................82.4气体传感器阵列........
本文标题:基于电子鼻的香油精快速检测模式识别及算法研究
链接地址:https://www.777doc.com/doc-69511 .html