A Real-Time Tracking System for Bread Production Based on YOLOv8 and DeepSORT

This article presents the application of bread production processes through integrating AI-driven detection and tracking systems. Utilizing the YOLOv8 and DeepSORT algorithms, this study explores developing and implementing a real-time system for detecting, counting, and reporting bread units. The methodology includes the preparation of a dataset from an operational bakery and subsequent training of the system on the Jetson Xavier NX embedded platform. Our findings show that the system can detect bread with a 96% accuracy rate during the training phase. Implementing this system in a bread production facility has significantly improved efficiency, waste reduction, and workflow optimization. By providing detailed insights into both the technical setup and the operational outcomes, this research underscores the potential of intelligent technologies to transform bread production.