RemoteSensingofLandscapeswithSpectralImages Spectralimages,especiallythosefromsatellitessuchasLandsat,areused worldwide for many purposes, ranging from monitoring environmental changes and evaluating natural resources to military operations. In a sig- nificantdeparturefromstandardremote-sensingtexts,thisbookdescribes how to process and interpret spectral images using physical models to bridge the gap between the engineering and theoretical sides of remote sensing and the world that we encounter when we put on our boots and ventureoutdoors. Remote Sensing of Landscapes with Spectral Images is designed as a textbookandreferenceforgraduatestudentsandprofessionalsinavariety ofdisciplinesincludingecology,forestry,geology,geography,urbanplan- ning,archeology,andcivilengineering,whowanttousespectralimagesto help solve problems in the field. The emphasis is on the practical use of images rather than on theory and mathematical derivations, although a knowledgeofcollege-levelphysicsisassumed.Examplesaredrawnfrom avarietyoflandscapesandinterpretationsaretestedagainsttherealityseen on the ground. The reader is led through analysis of real images (using figuresandexplanations),andtheexamplesarechosentoillustrateimpor- tant aspects of the analytic framework, rather than simply how specific algorithmswork. Thisbookissupplementedbyawebsitehostingdigitalcolorversionsof figuresinthebookaswellasancillarycolorimages(www.cambridge.org/ 9780521662215). Remote Sensing of Landscapes with Spectral Images A Physical Modeling Approach John B. Adams and Alan R. Gillespie DepartmentofEarthandSpace Sciences UniversityofWashington CAMBRIDGEUNIVERSITYPRESS Cambridge,NewYork,Melbourne,Madrid,CapeTown,Singapore,Sa˜oPaulo CAMBRIDGEUNIVERSITYPRESS TheEdinburghBuilding,Cambridge,CB22RU,UK PublishedintheUnitedStatesofAmericabyCambridgeUniversityPress,NewYork www.cambridge.org Informationonthistitle:www.cambridge.org/9780521662215 #CambridgeUniversityPress2006 Thispublicationisincopyright.Subjecttostatutoryexception andtotheprovisionsofrelevantcollectivelicensingagreements, noreproductionofanypartmaytakeplacewithout thewrittenpermissionofCambridgeUniversityPress. Firstpublished2006 PrintedintheUnitedKingdomattheUniversityPress,Cambridge AcatalogrecordforthispublicationisavailablefromtheBritishLibrary ISBN-13978-0-521-66221-5hardback ISBN-100-521-66221-4hardback CambridgeUniversityPresshasnoresponsibilityfor thepersistenceoraccuracyofURLsforexternalor third-partyinternetwebsitesreferredtointhispublication, anddoesnotguaranteethatanycontentonsuch websitesis,orwillremain,accurateorappropriate. ToCarylandKaren Contents Abouttheauthors pageix Preface xi Acknowledgments xiv 1 Extractinginformationfromspectralimages 1 1.1 Introduction 1 1.2 Fieldstudiesandspectralimages 3 1.3 Photointerpretationofspectralimages 7 1.4 Spectralanalysisofimages 19 1.5 Testingandvalidatingresults 27 1.6 Summarystepsforextractinginformation 35 2 Spectroscopyoflandscapes 39 2.1 Basicsofspectroscopyforfieldinvestigators 39 2.2 Spectroscopyatlandscapescales 52 2.3 Spectroscopyappliedtoimages 60 3 Standardmethodsforanalyzingspectralimages 65 3.1 Initialevaluation 65 3.2 Calibration 70 3.3 Enhancementforphotointerpretation 81 3.4 Datareconnaissanceandorganization 84 3.5 Physicalmodelingwithspectraldata 112 4 Spectral-mixtureanalysis 126 4.1 Endmembers,fractions,andresiduals 128 4.2 Shade 135 4.3 Fractionimages 137 4.4 Findingendmembers 145 4.5 Calibrationfeedback 159 4.6 Nonlinearmixing 164 4.7 Thermal-infraredimages 165 5 Fractionimagesoflandscapes 168 5.1 Whattodowithfractionimages 168 5.2 Classificationusingendmemberfractions 183 vii viii Contents 6 Targetdetection 192 6.1 Spectralcontrastandtargetdetection 192 6.2 Detectionlimits 224 6.3 Spectralcontrastandspatialscale 237 7 Thematicmappingoflandscapes 244 7.1 Fieldmapsandimage-derivedmaps 244 7.2 Thematicmappingwithspectralimages 250 8 Processesandchange 298 8.1 Processpathwaysinspectralimages 298 8.2 Referencepathways 312 8.3 Mappingchangesinlandscapes 324 Glossary 337 Reference 350 Index 357 About the authors JohnB.Adams JohnAdamsworkedonearlygeologicalexplorationoftheMoon,planets, andasteroids,andwasinstrumentalindevelopingreflectancespectroscopy asaremote-sensingmethodformineralidentification.Asamemberofthe scientific team that studied the first lunar samples, he demonstrated that Earth-basedtelescopicspectracouldbeusedtoidentifyandmaprocktypes on the Moon. He has used spectroscopy-based remote sensing of Earth tostudygeomorphicprocessesinarid regionsandtointerpretchanges in land use in temperate and tropical landscapes around the world. Adams establishedtheRemoteSensingLaboratoryattheUniversityofWashington inSeattlein1975.HepresentlyisEmeritusProfessorintheDepartmentof Earth and Space Sciences at the University of Washington. He spends as muchtimeaspossibleintheNorthCascadeMountainsofWashington. AlanR.Gillespie AlanGillespieisaProfessorofEarthandSpaceSciencesattheUniversity of Washington in Seattle, where he has been since 1987. He has been involved with remote sensing since joining the Mars Mariner project at Caltech’sJetPropulsionLaboratorylatein1969.WiththelaunchofERTS-1, Gillespie switched to terrestrial remote sensing, focusing first on image processingandthenonapplications,withemphasisonthethermalinfrared. He has been a member of the US Terra/ASTER team since 1991, and is responsible for the temperature/emissivity separation algorithms and standard products. Gillespie is also a glacial geologist with a strong interestinpaleoclimate.HegraduatedfromCaltechin1982withaPh.D. ingeologyandathesisonthe glacialhistory ofthe SierraNevada.Since 1991 he has been using remote sensing and field studies to elaborate the regionalvariationsintheglacialhistoryofcentralAsia.WithBarrySiegal, Gillespie contributed to and edited Remote Sensing in Geology (1980). GillespieiscurrentlytheeditorofQuaternaryResearch. ix