Risk Management of Personnel Intrusion at Container Terminals Based on Visual Detection

At the port terminal, the invasion of personnel into hazardous areas is a common safety hazard, often resulting in significant personnel and economic losses. In order to address this issue, researchers have developed an algorithm based on target detection. This algorithm combines target detection with a new IoU-p dynamic threshold strategy to achieve rapid and efficient detection of personnel intrusion in hazardous areas. We also discuss a strategy for detecting intrusion of distant targets. Experimental results based on a constructed benchmark dataset show that the proposed algorithm can effectively detect events of personnel intrusion into hazardous areas at container terminals and issue timely alarms.