04高光谱遥感第四讲

Lecture4RadiometricCalibrationofHyperspectralRemoteSensingData™™™™™™DN™2009™™™DN™™™™™™™™™™™™™™DN=AL+DN0DN=AL2+BL+DN0L=ADN+L0L=ADN2+BDN+L0LMAXDNMAX,LMINDNMINA=(LMAX-LMIN)/(DNMAX-DNMIN)CCBBCHBBB(THBB)B(TCBB)L=aDN+b™ii2121()()()iiLSLdWmsrλλλλλ−−=⋅⋅⋅∫)(λL)(λiS1λ2λiL())()()()(11200−−−∞∞⋅⋅⋅⋅=∫∫msrmWdSdLSLiieiμλλλλλiiieiADCDCL/)(,0−=iAiDCiDC,0LandsatTMMODISDN)()()()(1200,−−∞∞⋅⋅⋅=∫∫mmWdSdESEisieisμλλλλλ)(λsEseiseiiELdθπρcos,2*⋅⋅=sθd™™™™™™™™1.zzz2.zzz3.zz™™™™™™)(/)(12λλAA™™)(/)(12tAtA™™2121),,(/),(λλλλ≈tAtA™™™™™AVIRIS1000w™™™1030cm™3000™™™™™™™™™™™™AVHRR,HRV/HRVIR,VEGETATION,TM/ETM+,POLDER,AVHRR,HRV/HRVIR,VEGETATION,TM/ETM+,POLDER,ASTER,MISR,MODISASTER,MISR,MODIS™™MERISMERIS™™SeaWIFSSeaWIFSMODISMODISLandsatTMLandsatTM//ETM+ETM+SPOTHRVSPOTHRVHRVIRHRVIRVEGETATIONVEGETATIONTM5TM5SPOTHRVSPOTHRVHRVIRHRVIRSeaWIFSSeaWIFSMERISMERISMISRMISRMODISMODISETM+ETM+SeaWIFSSeaWIFSMERISMERISMISRMISRMODISMODISETM+ETM+™™DN™‡‡‡Li(1cos/1)22tppLETLTLλθπρ=++Eλ:θ:T1T2-Lp1-Lp2-ρ1.2.3.4.5.6.7.8.™999Spectraldimension246pixelsSpatialdimension328pixelsLongerwavelength•VerticalStripesCausedby¾DetectorNonlinearities()¾TemperatureEffects()¾MovementoftheslitwithrespecttothefocalPlane()¾Calibrationofsomedetectorswithsignificantlymodifiedgainandoffset()¾Disabilityofdetectors()•De-stripingApproaches()¾Interpolation¾Forward-inverseMNFTransform¾HistogramMatching(offset-gainadjustment)¾FourierTransform,wavelettransform,etc.StripeNoiseOMISDNSabbSaDN+⋅=iaiaaε=+ibibbε=+-biiaiσσε−=*iibiMM−=*εiiiiMMDNDN***)(+−=σσFourierTwoStripesBeforeAfterInterpolationBand13BeforeAfterForward-InverseMNFTransformationBand90HyperionData(MomentMatch)()()()referenceiadjustedinitialsamplerefenenceisampleDNDNuuσσ=−+():iiisampleureferenceuisampleσreferenceσ:(MomentMatch)()()refenenceisamplereferenceisampleinitialadjusteduuDNDN+−=σσ)(():(MomentMatch)(PHI200080373)FlatField=R/FRFRailroadValleyAnexampleofanimagebeforeandafterflatfieldcalibrationBeforeAfterIARR=R/FRFELReflectance=gainradiance+offset(errors)123DNb=().b+bb:b:™™N1()2()3()...n()ABMTF……