The demand for AI-based real-time monitoring is rapidly expanding in the field of industrial safety management. Particularly in large-scale infrastructure environments such as airports, railways, and logistics, the importance of 'Edge AI' technology—which detects dangerous situations immediately on-site without network delays—is growing.
SAIGE, an industrial AX company, won the grand prize in the AI startup category of the 'Incheon Airport AI-PORT Idea Contest' organized by Incheon International Airport Corporation.
Sage was selected as the best idea in the field of airport shuttle train safety control for proposing the 'Edge AI-based Integrated Safety Control System for Shuttle Trains.' With this award, the company will also proceed with a Proof of Concept (PoC) at Incheon Airport.
Edge AI applied to airport safety control
The system proposed by Sage involves connecting an AI semiconductor (NPU)-based appliance to the existing CCTV infrastructure. This enables real-time processing at the camera level for functions such as detecting intrusions into hazardous high-voltage power line zones, identifying non-wearing of personal protective equipment (PPE), and detecting worker collapses.
Existing server-based video analysis methods had limitations in responding to dangerous situations due to the need for network transmission and server computation. In contrast, Sage's Edge NPU architecture performs AI computations directly on-site, enabling immediate risk detection without communication delays.
The company explained that it can operate independently, particularly in communication dead zones such as underground tunnels or substations, and offers advantages in terms of complying with airport security regulations as it does not transmit original video data externally.
The Incheon Airport Shuttle Train is a driverless automatic (UTO) light rail system connecting Terminal 1, the boarding area, and Terminal 2, serving as a key transportation infrastructure used by over 20 million people annually. Currently, the AI CCTVs in operation are primarily utilized for analyzing platform congestion, leaving areas such as high-voltage line intrusion and worker safety detection as relatively blind spots.
Sage plans to verify the intrusion detection accuracy, fall detection response speed, and helmet identification accuracy for five platform doors during the PoC phase, and then expand the scope of application in the future.
Recently, competition in industrial settings has intensified not only for generative AI but also for 'physical AI' and edge AI technologies that operate immediately in real-world environments. In particular, demand for field-oriented AI for real-time safety response and operational efficiency is rapidly increasing in the manufacturing, transportation, and public infrastructure sectors.
Lee Seong-woo, Marketing Director at Sage, stated, “This award is a recognition that Sage’s Edge AI safety technology is a solution immediately applicable even in complex infrastructure environments like airports,” adding, “Starting with the Incheon Airport PoC, we plan to expand AI safety control to various transportation infrastructure sectors, including railways and urban rail.”
SAIGE operates 'SAIGE SAFETY,' an AI solution that detects types of industrial safety accidents—such as fires, smoke, collapses, and failure to wear protective equipment—at manufacturing, construction, and port sites. With this airport demonstration, the company plans to expand its business into the transportation infrastructure sector.