基于數(shù)字圖像處理的儲糧害蟲圖像邊緣檢測研究.pdf

1、“∞q一——＆f℃nL叫直L’‘IJⅡⅡ∞凹—～＆～辛卞斜技大署碩士學(xué)位論文基于數(shù)字圖像處理的儲糧害蟲圖像邊緣檢測研究學(xué)位申請人學(xué)科專業(yè)指導(dǎo)教師眷辯丌期歲濤華水利水電上槲周建中教授200511f0華中科技大學(xué)碩士學(xué)位論文j；；■=目==z=====■■●■■__目；《=====t===；=■■●_=；；======E■■■■■■●■■t；；==========；■■■目1目=z=j■E■=■_■皇AbstractChinaisthela

2、rgestcountryofgrainproduction，storageandconsumptionintheworldDoingwellstored—grainmanagementisaveryimportanttilingabouttilenationaleconomyandthepeople’SlivelihoodInrecentyears，thestoredgrainismorethan500billionsofkilogra

3、msinourcountrynecentergovernmentoffersbillionsofRMBtothegraIndeportsinordertomanagingstoragewellButplentyofstoragestillwasattaIntedbecauseoftheillmanagementThestored—grainlossisveryseriouscalculatedbythecurrentO2％lossrat

4、iointhenationalgraindeportsThestored—graininsectpestisoneoftheimportantfacts’7rileFifthGrainTradeScienceandTechnologyDevelopmentProgramming”putforwarddefinitelytorealizeautomaticdetectionaboutstored—graininsectpestsThewa

5、ysoftheSampling，theSoundDetecting，theNearInfraredandothersinthewordcan’tsupplywellandtrulythegrainpestscategory，denseandotherparametersInaddition，withtheincreaseofthestored—grainpestsdrugfastness，theircategoryanddenseare

6、increasinginrecentyearsAsaresult，developingakindofscientific，precision，simpledetectiontechnologyforstoredgrainpestsisverynecessaryandimperiousThereareaseriesofadvantagesmakinguseoftheimagerecognitiontechnologytodetecttil

7、estored—grainpests，suchashighprecise，lowprice，highcfficiencynopollution，lesslabor，convenientconnectionwiththecomputergraindetectioningraindeports，andSOonItCanhelptilegrainmanagerstomakescientificdecisions；inordertotheyCa

8、ntakerationalpreventionmeaSuresintime，thestoragecanbemanagedinqualityquantityandgreennessThemainWjrkiSasfollows：Namelyutilizecomputervisionpatternprocessandrecognitiontechnologyofthemode，carryonthemeaSuringofthegrainworl

9、n，judgeanddiscernForgathergoodpicture，grainwormandgrainmixtogetherinordertodiscempests，Wecarryontheedgetomeasuretothepictureatfirst，bymeasuringtheedgeoutlinewhichgetsoffthegrainwormCarryontoitthresholdvaluecutapart，inord