Design andapplication of a UAV-based intelligent inspection system for blast furnaces

To address the issues of insufficient coverage and delayed response of manual inspection in blast furnace ironmaking areas, this paper presents a UAV-based intelligent inspection system that enables real-time monitoring and automated inspection of ironmaking areas. Based on industrial-grade UAVs and fully automated docking stations, the system features a four-tier hardware architecture and a layered software architecture, integrating temperature monitoring, deflection monitoring and anomaly alarming. At the algorithmic level, an infrared-radiation-to-temperature mapping model and a composite alarming scheme with operation-adaptive thresholds are established, and a machine-vision-based pixel-level method is proposed for measuring pipeline deflection. Field tests show that the UAV-based infrared temperature measurement error remains below 10 ℃, the maximum relative error of downcomer deflection measurement is 3.11%, and the system uptime reaches 100%, demonstrating the feasibility and effectiveness of the proposed approach for the intelligent operation and maintenance of blast furnaces.