A novel campus fire monitoring system based on YOLOv5

To protect lives and equipment, a robust, reliable, and expandable fire monitoring system is critical to detect and react in a campus fire situation. Thus, this paper proposes a YOLOv5-based campus fire monitoring system directly deployable on existing terminals, computers, and other IoT terminals. The proposed campus fire monitoring system effectively monitoring fires in various scenarios on campus all day long. Our deep learning model has been trained using fire images and fire scene videos. The experimental evaluations on different end devices demonstrate that our system achieves an appealing performance in campus fire monitoring.