117. Sugar beet disease detection based on remote sensing data and artificial intelligence
Résumé
Plant disease detection for precision chemicals application is a field of interest to reduce the use of chemicals in the field. Detection of diseases with an RGB sensor camera simulating real acquisition on a sprayer was performed. For dataset preparation, about 986 images were collected from a sugar beet field in North France under varying conditions. The accuracy assessment was performed using statistical indicators which are Precision, Recall, F1-score, and Average Precision. The Mask RCNN model detects sugar beet disease with an average precision of 92.73 % for the three disease types. Artificial intelligence has a strong potential to reduce the use of chemicals thus costs optimization and environmental risks minimization over agricultural fields.
Origine : Fichiers produits par l'(les) auteur(s)