Overview: On November 1, VnExpress reported that Singapore Changi Airport is collaborating with local technology company NCS to upgrade its AI-based Foreign Object Debris (FOD) monitoring system. The goal is to identify objects as small as 4 centimeters with an accuracy rate of approximately 95%, according to Singapore’s Lianhe Zaobao. This system, known as iFerret, is transforming runway inspections from “scheduled manual inspections” to “24/7 machine monitoring.”
According to public reports, iFerret has deployed a network of high-definition cameras along the runway to continuously scan the runway surface. An AI algorithm is used to distinguish actual foreign objects from "visual interference" such as shadows and reflections from wet runway surfaces. Unlike systems that rely on regular manual inspections, this system can operate 24/7. Once a suspicious object is detected, airport personnel can locate and address it more quickly, reducing the time required to close the runway for inspection.
In the aviation industry, metal fragments, detached parts, and other objects that could potentially damage aircraft or interfere with takeoffs and landings are collectively referred to as Foreign Object Debris (FOD). Historically, the 2000 Concorde crash was linked to debris on the runway, and since then, airports worldwide have significantly increased their investment in FOD detection. Small-sized debris is precisely the type that is most difficult for human inspectors to detect promptly, which is where the value of machine vision lies.
According to VnExpress, citing data from NCS and the Straits Times, the system's initial version was introduced in 2009, and a more advanced version was deployed starting in 2021, reducing the number of false alarms by approximately 90%. Most false alarms occur in situations such as rainy days with reflections. Reducing false alarms helps prevent inspection personnel from being diverted by irrelevant alerts, allowing them to focus on genuine threats.
This round of collaboration focuses on improving the performance and expanding the deployment of the third-generation system. As of this writing, NCS has not announced the completion date for the third-generation version. In other words, this is a strategy of "continuous upgrades to existing systems" rather than a one-time equipment replacement.
In terms of background, Changi Airport was ranked as the World’s Best Airport in the 2026 Skytrax World Airport Awards, placing it ahead of Incheon in South Korea and Haneda in Japan. Furthermore, Changi Airport’s Terminal 5 is expected to be operational in the mid-2030s, and the anticipated increase in passenger volume and aircraft movements will place even greater demands on runway availability and safety. By assigning AI to handle the high-frequency, repetitive task of runway monitoring, it effectively adds an additional layer of non-human redundancy to ensure runway safety.
It is worth noting that runway AI monitoring is not an isolated application. Several major hub airports around the world have already deployed similar capabilities in areas such as biometric immigration control, AI-powered flight delay prediction, and apron safety inspections. The "perception-interpretation-action" closed-loop system is expanding from individual scenarios to the entire airport area.
Runway safety is the most fundamental safety requirement in civil aviation. The significance of this upgrade lies not in the use of AI, but in transforming the process from "humans identifying potential hazards" to "systems identifying potential hazards and humans making decisions" – delegating hazard identification to machines and leaving the decision-making process to humans. For hub airports with near-saturated throughput, these seemingly minor but highly valuable AI applications may be more worthy of emulation than the "smart services" that passengers can directly perceive.