Automated Detection of Pineapple Plants in Drone-Captured Aerial Imagery for Precision Agriculture
Year: 2024
Type: conference-paper
Abstract: Pineapple harvesting remains largely manual due to scattered planting patterns and complex fruit structure. This study presents a method for detecting pineapple plants in large orthomosaic images using the Slicing Aided Hyper Inference (SAHI) technique combined with the YOLOv8 segmentation model. SAHI divides large images into smaller patches, enabling accurate detection. A dataset of 867 training and 97 validation images from two orthomosaics was used, with the model achieving 93% precision and 88% accuracy. Despite high precision, chal lenges with false negatives suggest future improvements. This approach shows promise for automating pineapple harvesting and improving agricultural efficiency.
Authors: Matheus Arroyo de Lima, Thiago Mantovani Martins, Vitor Matheus Soares Siqueira De Almeida, Rafael Buosi Jardim, Victor Hugo Zani, Luiza Hoehl Loureiro Alves Barbosa
Institutions: Centre de Biophysique Moléculaire, Universidade de São Paulo
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
SDG: Zero hunger
Open Access status: gold
OpenAlex ID: W4408869774