Akademik Veri Yönetim Sistemi
31st International Conference on Artificial Life and Robotics, ICAROB 2026, Oita, Japonya, 29 Ocak - 01 Şubat 2026, ss.206-210, (Tam Metin Bildiri)
Automated tomato detection is essential for robotic harvesting and reducing labor dependency in agriculture. However, robust tomato detection under real farming conditions remains a critical challenge in computer vision. This study aimsto develop a reliable system that can accurately recognize tomatoes in images and clearly annotate their locations. Farmers rely on automation to help automate their harvesting operations and reduce the time spent on manual labor. Tomatoes present a difficult test case because they vary widely in color, size, shape, and ripeness. Therefore, selected tomatoes were harvested specifically to evaluate whether current AI models can handle these real-world variations. YOLOv5 and YOLOv8 were compared to evaluate their performance across various real-time object detection algorithms. Overall, the tomato detection framework demonstrates that machine learning can significantly enhance agricultural efficiency and support more informed farm management.