Close-rangehyperspectralimagesareapromisingsourceofinformationinplantbiology,inparticular,forinvivostudyofphysiologicalchanges.Inthisstudy,weinvestigatehowdatafusioncanimprovethedetectionofleafelementsbycombiningpixelreflectanceandmorphologicalinformation.Thedetectionofimageregionsassociatedtotheleafstructuresisthefirststeptowardquantitativeanalysisonthephysicaleffectsthatgeneticmanipulation,diseaseinfections,andenvironmentalconditionshaveinplants.WetestedourfusionapproachonMusaacuminata(banana)leafimagesandcompareditsdiscriminantcapabilitytosimilartechniquesusedinremotesensing.Experimentalresultsdemonstratetheefficiencyofourfusionapproach,withsignificantimprovementsoversomeconventionalmethods.
