Lightweight Intelligent Fire Monitoring System Based on Flask and YOLOv8

In recent years, with the impact of global warming, the incidence of fire has increased, and the prevention and control of fire has become an important topic of safety management. Traditional fire alarm systems rely on temperature and smoke sensors, but they often experience delays, making early fire detection challenging. Based on this requirement, this system aims to realize real-time fire monitoring and provide more efficient early fire warning by introducing computer vision technology. The system uses the improved YOLOv8 lightweight model as the core fire detection algorithm, and realizes the front-end transmission and display of video streams through the Web service constructed by the lightweight Flask framework. The system includes three core modules: front-end page, back-end server and model reasoning, to realize the automation of the whole process from video capture, image analysis to result display. Compared with YOLOv8n, the recall rate of our model is increased by 1.9% and the mAP is increased by 0.5%, while the model size is reduced by 15.4% and the calculation amount is reduced by 21.0%. In general, the system has a lightweight structure and fast reasoning, which ensures the real-time response ability of the system, and is also convenient for flexible deployment on various hardware environments.