高分五号02星高光谱影像典型植被结构和生理参数反演
II
北华航天工业学院硕士学位论文
Abstract
Vegetationisanimportantcomponentofterrestrialecosystems,andvariousdynamic
processesofvegetationarecloselyrelatedtovegetationparameters.Therefore,accurate
inversionoftypicalphysiologicalparametersofvegetationiscrucialformonitoring
vegetationgrowthstatus.Theremotesensinginversionmethodsforsurfacevegetation
parametersaremostlybasedontheTopofcanopy(TOC)reflectance,butthesatellitedirectly
observestheTopofatmosphere(TOA)reflectance,andtheatmosphericcorrectionwhen
obtainingTOCreflectanceusuallyassumesthatthesurfaceisalambertsurface,ignoringthe
anisotropyinformationofvegetationcanopy.Therefore,thisstudyisbasedontheSoilPlant
AtmosphereRadiationTransferModel(SPART)toexplorethefeasibilityofinvertingtypical
vegetationparametersbasedonTOAreflectance.
ThisstudyusesthehyperspectralimagesofGaofen-502satelliteasthedatasource,with
soybeanandcorncropsastheresearchobjects,toinvertandverifytheaccuracyoftypical
vegetationstructuralparameters-leafareaindex(LAI)andphysiologicalparameters-leaf
chlorophyllcontent(Cab).Themainresearchcontentofthepaperisasfollows:
(1)Aimingattheproblemsofmanyparametersinthephysicalradiativetransfermodel
andtheredundancyofhyperspectralbands,weconductedaglobalsensitivityanalysisofthe
PROSAILmodeltorevealtheinteractionbetweenvariousphysiologicalparametersof
vegetationandquantitativelyanalyzethesensitivebandsofLAIandCab.Combinedwiththe
responseofLAIandCabtoTOCreflectance,aphasedinversionstrategywasdeveloped.
(2)Wes