Ball Detection and Color Identification for a Mobile Robot using a 2D Camera

In this study, a novel method is developed to help the mobile robot system accurately detect and recognize the color of a ball in environments with light disturbances using deep learning. The YOLOv8 algorithm is applied to detect the ball and identify its color. The effectiveness of the algorithm is tested in various lighting conditions and when the balls are inside a silo and when they are outside. The developed algorithm identifies balls even when they are partially obscured by shadows.