基于高光譜紅邊參數(shù)定量提取棉花冠層特征信息的研究.pdf

1、石河子大學(xué)碩士學(xué)位論文基于高光譜紅邊參數(shù)定量提取棉花冠層特征信息的研究姓名：袁杰申請學(xué)位級別：碩士專業(yè)：作物栽培學(xué)與耕作學(xué)指導(dǎo)教師：王登偉20070501基于高光謙紅邊參數(shù)定量提取棉花冠層特征倍息的研究ResearchofCropCanopyCharacteristicInformationbyUsingHyperspectralRedEdgeParametersofCottonAbstractInthisstudyweselectco

2、ttonasourinteres血gobjectanduseASDHeldSpecProFRtoobtafnthespedctralreflectanceofcoRonWecaIriedfieldspectralobservationsatecottondifferentgrowthperiodandobtainedspectra]reflectanceofcottoncanopyBesides，thecorrespondingcano

3、pycharacteristicinformationweremeasuredatthesametimeThemajor啪oPYcharacleristic礎(chǔ)ormationincladeL址MLATCDPTCRPKMLD，AFMandADMExceisoftwareandSPSS(StatisticalpackageforSocial)areusedtoprocessdataMultifactorstatisticstechnique

4、andcurvefitnessanalysisareadoptedtodiscussthecorrelationbetweencottoncanopycharacteristicinformationandrededgeparametersaswellasitsdifferenttransformations，whichincludeitsfirstderivationspectralreflectance，rededge，hypers

5、poctralcharacteristicvariablesvegetationindicesAndthenthehyperspectralremotesensingestimationmodelsfordi缸renteaooPYcharacteristicinformationarecom廿uctedbasedontheresnItsofcorrelationanalysis，finallythepredictiveprecision

6、sareanalyzedforsomechosenmodelsinordertodeterminethebestestimationmodelsforeverycanopycharacteristicinformationThccorrelationanalysisofcanopycharacteristicinformationandrededgeparameterswerepresentTheresultsindicated：the

7、Depthi，Ateai，NDiandEGFNcouldbeusedtoestimateLAJandAFM；SRoforLAI，MLAandK；dkforLAJ，MU)andADM；LwidthforLAJandAFMThereaftertheregressionmodelsofretrievingcanopycharacteristicinformationbasedOnrededgeparametersobtainablefromr

8、emotelysensingdata，werealsoestablishedrespectivelyRededgeparameterscouldbeconsideredasasensitiveindicatorforcottonnutritionstatusbynondestructinomeansonalargescaleWeanalysetheregularpatternofrededgealongwiththedifferentg

9、rowthstagesTheresultsshowedthatthesignificantlycorrelationbetweenNDVIandcottoncanopycover(r=06303，n；62)SowecallusingNDVItoestimatecottoncanopycoverTheslopeofrededgequadraticmodelhasthehighestcorrelationtoe伍cient(r=O6983，

10、o=62)ItisfcasibleforusingRMSEofestractingcottoncanopycover(RMSE004769／mz)1hestudyindicatesthatitcanplayavitalroleinprobidingtimespecificandtimecriticalinformationbyusinghyperspectralremotesensingforprecis／onfarmingForcot

11、tonatdifferentgrowthstages，conductedstatisticalcorrelation柚alysisbetweenvariouscanopycharacteristicinformationand43spectraparameters。detectedouttheoptimalspectralparameterstomakeasoundfoundationforsetupmodelsbyusingspect

12、ralparametersincludingstatisticalcorrelationanalysisonplantMLA、TCDP、TCRP、K、MID、LAj、AFM、ADMandcano口Yspectralreflectancespe斌raSetupstatisticalregressionmodelforvariouscottoncanopycharacteristicinformation，basedonspectralpa