Research on Intelligent Operation Safety Management Systems for Civil Aviation Airports Based on Deep Learning

As airport operations become increasingly complex, there is a growing demand for advanced security systems capable of real-time monitoring and management of human behavior. This study introduces an intelligent security and behavior management system tailored for airports, utilizing the YOLOv5 model as the primary recognition algorithm. The system captures and processes live video streams from strategically placed cameras, employing YOLOv5 for accurate detection, classification, and localization of individuals within the monitored area. Key features include reflective clothing detection and electronic fence monitoring, which works to ensure the precise identification of authorized personnel and the triggering of alarms for boundary violations. The system also incorporates a geo-check mechanism to verify the location of detected individuals within designated safety zones. Extensive testing was conducted using real-time camera feeds, scenariospecific video inferences, and arbitrary local videos to evaluate the model’s performance in diverse conditions. The results demonstrate the system’s robust ability to enhance airport security by providing timely and reliable data for effective management, positioning it as a valuable tool for modern airport safety operations.