Series ISSN: 1559-8136 A Z Synthesis Lectures on A R A N Image, Video & Multimedia Processing G • K E H T A R N Series Editor: Alan C. Bovik, University of Texas, Austin A V A Z Image Fusion in Image Fusion in Remote Sensing Conventional and Deep Learning Approaches Arian Azarang, The University of Texas at Dallas Remote Sensing Nasser Kehtarnavaz, The University of Texas at Dallas I(mmuaglteis fpuesciotrna li)n i mremagoetse t sheants ainreg coarp ptuanresdh abryp deniffinegre inntv soelnvesso rfus soinn gs astpealtliitael s(.p Thanicsh brooomka atidcd) raenssde ssp iemctargael I Conventional and Deep M fusion approaches for remote sensing applications. Both conventional and deep learning approaches A G are covered. First, the conventional approaches to image fusion in remote sensing are discussed. E F These approaches include component substitution, multi-resolution, and model-based algorithms. U S I Then, the recently developed deep learning approaches involving single-objective and multi-objective O N Learning Approaches loss functions are discussed. Experimental results are provided comparing conventional and deep IN R learning approaches in terms of both low-resolution and full-resolution objective metrics that are E M commonly used in remote sensing. The book is concluded by stating anticipated future trends in O T pansharpening or image fusion in remote sensing. E S E N S I N G Arian Azarang Nasser Kehtarnavaz About SYNTHESIS This volume is a printed version of a work that appears in the Synthesis Digital Library of Engineering and Computer Science. Synthesis books provide concise, original presentations of important research and M O development topics, published quickly, in digital and print formats. R G A N Synthesis Lectures on & C Image, Video & Multimedia Processing L A Y store.morganclaypool.com P O O L Image Fusion in Remote Sensing Conventional and Deep Learning Approaches Synthesis Lectures on Image, Video, and Multimedia Processing Editor AlanC.Bovik,UniversityofTexas,Austin TheLecturesonImage,VideoandMultimediaProcessingareintendedtoprovideauniqueand groundbreakingforumfortheworld’sexpertsinthefieldtoexpresstheirknowledgeinuniqueand effectiveways.ItisourintentionthattheSerieswillcontainLecturesofbasic,intermediate,and advancedmaterialdependingonthetopicalmatterandtheauthors’levelofdiscourse.Itisalso intendedthattheseLecturesdepartfromtheusualdrytextbookformatandinsteadgivetheauthor theopportunitytospeakmoredirectlytothereader,andtounfoldthesubjectmatterfromamore personalpointofview.Thesuccessofthiscandidapproachtotechnicalwritingwillrestonour selectionofexceptionallydistinguishedauthors,whohavebeenchosenfortheirnoteworthy leadershipindevelopingnewideasinimage,video,andmultimediaprocessingresearch, development,andeducation. Intermsofthesubjectmatterfortheseries,therearefewlimitationsthatwewillimposeother thantheLecturesberelatedtoaspectsoftheimagingsciencesthatarerelevanttofurtheringour understandingoftheprocessesbywhichimages,videos,andmultimediasignalsareformed, processedforvarioustasks,andperceivedbyhumanviewers.Thesecategoriesarenaturallyquite broad,fortworeasons:First,measuring,processing,andunderstandingperceptualsignalsinvolves broadcategoriesofscientificinquiry,includingoptics,surfacephysics,visualpsychophysicsand neurophysiology,informationtheory,computergraphics,displayandprintingtechnology,artificial intelligence,neuralnetworks,harmonicanalysis,andsoon.Secondly,thedomainofapplicationof thesemethodsislimitedonlybythenumberofbranchesofscience,engineering,andindustrythat utilizeaudio,visual,andotherperceptualsignalstoconveyinformation.Weanticipatethatthe Lecturesinthisserieswilldramaticallyinfluencefuturethoughtonthesesubjectsasthe Twenty-FirstCenturyunfolds. ImageFusioninRemoteSensing:ConventionalandDeepLearningApproaches ArianAzarangandNasserKehtarnavaz 2021 MultimodalLearningTowardMicro-VideoUnderstanding LigiangNie,MengLiu.andXuemengSong 2019 iv VirtualRealityandVirtualEnvironmentsin10Lectures StanislavStanković 2015 DictionaryLearninginVisualComputing QiangZhangandBaoxinLi 2015 CombatingBadWeatherPartII:FogRemovalfromImageandVideo SudiptaMukhopadhyayandAbhishekKumarTripathi 2015 CombatingBadWeatherPartI:RainRemovalfromVideo SudiptaMukhopadhyayandAbhishekKumarTripathi 2014 ImageUnderstandingUsingSparseRepresentations JayaramanJ.Thiagarajan,KarthikeyanNatesanRamamurthy,PavanTuraga,andAndreasSpanias 2014 ContextualAnalysisofVideos MyoThida,How-lungEng,DorothyMonekosso,andPaoloRemagnino 2013 WaveletImageCompression WilliamA.Pearlman 2013 RemoteSensingImageProcessing GustavoCamps-Valls,DevisTuia,LuisGómez-Chova,SandraJiménez,andJesúsMalo 2011 TheStructureandPropertiesofColorSpacesandtheRepresentationofColorImages EricDubois 2009 BiomedicalImageAnalysis:Segmentation ScottT.ActonandNilanjanRay 2009 JointSource-ChannelVideoTransmission FanZhaiandAggelosKatsaggelos 2007 SuperResolutionofImagesandVideo AggelosK.Katsaggelos,RafaelMolina,andJavierMateos 2007 v TensorVoting:APerceptualOrganizationApproachtoComputerVisionandMachine Learning PhilipposMordohaiandGérardMedioni 2006 LightFieldSampling ChaZhangandTsuhanChen 2006 Real-TimeImageandVideoProcessing:FromResearchtoReality NasserKehtarnavazandMarkGamadia 2006 MPEG-4BeyondConventionalVideoCoding:ObjectCoding,Resilience,and Scalability MihaelavanderSchaar,DeepakSTuraga,andThomasStockhammer 2006 ModernImageQualityAssessment ZhouWangandAlanC.Bovik 2006 BiomedicalImageAnalysis:Tracking ScottT.ActonandNilanjanRay 2006 RecognitionofHumansandTheirActivitiesUsingVideo RamaChellappa,AmitK.Roy-Chowdhury,andS.KevinZhou 2005 Copyright©2021byMorgan&Claypool Allrightsreserved.Nopartofthispublicationmaybereproduced,storedinaretrievalsystem,ortransmittedin anyformorbyanymeans—electronic,mechanical,photocopy,recording,oranyotherexceptforbriefquotations inprintedreviews,withoutthepriorpermissionofthepublisher. ImageFusioninRemoteSensing:ConventionalandDeepLearningApproaches ArianAzarangandNasserKehtarnavaz www.morganclaypool.com ISBN:9781636390741 paperback ISBN:9781636390758 ebook ISBN:9781636390765 hardcover DOI10.2200/S01074ED1V01Y202101IVM021 APublicationintheMorgan&ClaypoolPublishersseries SYNTHESISLECTURESONIMAGE,VIDEO,ANDMULTIMEDIAPROCESSING Lecture#13 SeriesEditor:AlanC.Bovik,UniversityofTexas,Austin SeriesISSN Print1559-8136 Electronic1559-8144 Image Fusion in Remote Sensing Conventional and Deep Learning Approaches Arian Azarang and Nasser Kehtarnavaz TheUniversityofTexasatDallas SYNTHESISLECTURESONIMAGE,VIDEO,ANDMULTIMEDIA PROCESSING#13 M &C Morgan &cLaypool publishers ABSTRACT Image fusion in remote sensing or pansharpening involves fusing spatial (panchromatic) and spectral (multispectral) images that are captured by different sensors on satellites. This book addresses image fusion approaches for remote sensing applications. Both conventional and deeplearningapproachesarecovered.First,theconventionalapproachestoimagefusioninre- motesensingarediscussed.Theseapproachesincludecomponentsubstitution,multi-resolution, andmodel-basedalgorithms.Then,therecentlydevelopeddeeplearningapproachesinvolving single-objective and multi-objective loss functions are discussed. Experimental results are pro- videdcomparingconventionalanddeeplearningapproachesintermsofbothlow-resolutionand full-resolution objective metrics that are commonly used in remote sensing. The book is con- cludedbystatinganticipatedfuturetrendsinpansharpeningorimagefusioninremotesensing. KEYWORDS imagefusioninremotesensing,pansharpening,deeplearning-basedimagefusion, fusionofspatialandspectralsatelliteimages