Incheon Airport develops AI system to detect baggage anomalies

Incheon International Airport Corporation and POSCO DX have co-developed an ‘AI-based baggage anomaly detection system’ and completed a patent application for the technology. The system utilizes deep learning models to analyze vision camera footage in real time, monitoring the shape, size, color, and loading patterns of luggage on conveyor belts.

The AI is designed to detect specific issues that could lead to misclassification or damage, such as ‘double loading’ (where two bags overlap and obstruct barcode recognition), extended suitcase handles that might snag on facilities, and luggage deviating from its transport tub. When an anomaly is detected, the system automatically redirects the item to a manual processing area for inspection.

During pilot testing at Incheon International Airport’s Terminal 1, the system demonstrated a 100% success rate in identifying anomalous baggage that could cause errors or damage. The companies are currently developing an additional detection system specifically for oversized luggage to further improve processing accuracy.