Cherry fruitlet detection using YOLOv5 or YOLOv8?
2024
Apeinans, Ilmars | Sondors, Marks | Litavniece, Lienīte | Kodors, Sergejs | Zarembo, Imants | Feldmane, Daina
Agriculture 5.0 incorporates autonomous decision-making systems in order to make agriculture more productive. Our study is related to the development of the autonomous orchard monitoring system using unnamed aerial vehicles for automatic fruiting assessment and yield forecasting. Respectively, artificial intelligence must be developed to count fruits in an orchard. The modern solutions are mainly data-based. Therefore, we collected and annotated cherry dataset with natural images (CherryBBCH81) for neural network training. The goal of the experiment was to select the optimal “You Look Only Once” (YOLO) model for the rapid development of fruit detection. Our experiment showed that YOLOv5m provided better results for CherryBBCH81 – mean average precision (mAP) at 0.5 0.886 in comparison with YOLOv8m [email protected] 0.870. However, additional tests with dataset Pear640 showed that YOLOv8m can outperform YOLOv5m: 0.951 vs 0.943 ([email protected]).
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Editeur Rēzeknes Tehnoloģiju akadēmija
ISSN 2256-070XCette notice bibliographique a été fournie par Fundamental Library of Latvia University of Life Sciences and Technologies
Découvrez la collection de ce fournisseur de données dans AGRIS