Monitoring and early warning method for material fire accidents in the production process based on thermal infrared imaging and object detection

Production equipment frequently operates under extreme conditions, such as high temperature, high pressure, and high speed, which present significant safety risks, including material leakage, fires, and explosions. Such accidents can result in considerable property loss, environmental damage, casualties, and adverse social impacts. Therefore, real-time monitoring and early warning of the production process are essential to promptly address potential hazards, prevent accidents, and ensure production safety. To reduce the likelihood of misjudgments and enhance early warning capabilities, this paper proposes a monitoring and early warning method for material fire accidents in the production process by integrating thermal infrared cameras with an object detection model. Utilizing thermal infrared imaging and temperature analysis technology, the system monitors temperature changes in the production area in real time, trains and develops a flame identification and detection model, and identifies and locates flames within the production area. A comprehensive flame monitoring and early warning method for production materials was developed, forming the basis of a monitoring and early warning system that facilitates real-time monitoring and prevention of material fires. The system visualizes temperature data trends using data visualization technology and combines this with on-site temperature data collection to enable comprehensive judgments, significantly enhancing monitoring and early warning efficiency. To evaluate the proposed method's effectiveness, the system was applied to an acrylic acid fractionation production unit for real-time flame monitoring and early warning. Experimental results demonstrate that the proposed method accurately monitors and provides early warnings for flames during the production process, offering an effective solution for preventing material fire accidents and possessing substantial engineering application value.