AA・Efficientfruitselectionbynon-destructiveevaluationoffruitqualityinpost-harvestmanagement・Phenotypingofdataanddeeplearning・Quantitativeanalysisoforchardsusingimagedataforhighqualityfruitproduction20YutaroOSAKOAssistantProfessor, Ph. DResearch keywords:・Fruitand fruit tree・Nondestructiveevaluation・Image phenotyping・Deep learning“Pomology”isabranchofhorticulturalsciencethatfocusesonfruitandfruittrees.Thisfieldinvolvesbasicresearchintophysiologyoffruitandfruittrees,aswellasappliedresearchnecessaryfortheefficientfruitproduction.Inparticular,ourresearchaimstodevelopnewapproachestoefficientfruitproduction.Wearegoingtodevelopnon-destructivephenotypingtechniquesusingdeeplearning(AI:ArtificialIntelligence),basedonfruitandorchardimages.Wealsoaimtoquantifyandanalyzecultivationtechniquesthatenablethehighqualityfruitproductionimagedataanalysis.fruittraitsusingimageAnexampleofamodelfordiscriminatingbetweensmall-seededfruitandlarge-seededfruit.Althoughthereisnodifferenceinappearancebetweenthetwofruits,deeplearningmodelsuggeststhatthefeaturesmaybepresentinthefruitapexandpeduncleside.Aexampleofadeeplearningmodelthatcaneasilydetectpersimmonfruitinimagesandvideos.Itcanleadtoyieldpredictionandmoreefficientcultivationmanagement.Bioresource ScienceDivisionLaboratory of Pomology
元のページ ../index.html#24