AI System Detects Man Overboard Incidents in Seconds

A new artificial intelligence system called ZOE is designed to detect passengers falling overboard from cruise ships, offering near-instantaneous alerts to crew members. Developed by the UK-based company Zelim, the technology aims to drastically improve survival rates in man-overboard situations, which are often fatal due to delayed detection.

ZOE can identify both adults and children falling into the water within four seconds, regardless of lighting conditions. The system provides an immediate playback loop of the incident, crucial for understanding the timing and circumstances. Currently, it can take hours to realize a person has gone overboard, by which time the potential search area can be vast, and the chances of survival significantly diminish. According to the World Health Organization, there is a 20% chance of retrieval within the first 25 minutes, but this drops to nearly 0% after an hour.

Tests conducted by Zelim with the ZOE device reportedly resulted in a nearly 100% rescue rate. Mike Collier, vice president of cruise at Zelim, stated that speed and certainty are paramount in man-overboard scenarios, and the system provides operators with reliable alerts to maximize rescue opportunities. He emphasized that undetected incidents can have severe consequences for passengers, crew, and rescue services.

Man-overboard incidents are rare, with an estimated 21 out of 30 million cruise passengers annually falling overboard. However, they are frequently fatal. Incidents often occur at night, when visibility is limited, making automated detection critical. "Survival depends on immediate detection," said Collier, noting that crews typically have under 11 minutes to respond, and in adverse conditions, as little as four to five minutes.

ZOE operates using advanced computer vision and video analytics, continuously monitoring the vessel's surroundings via a network of optical and thermal cameras. It automatically recognizes when a person enters the water and alerts the bridge within seconds. The system is designed to function in various weather conditions, including rain, snow, and rough seas, ensuring continuous surveillance. Crucially, ZOE can also track an individual in the water as the ship moves away, providing precise location information throughout the rescue operation. This continuous tracking helps reduce search times and increases the likelihood of a successful recovery.

Zelim tested ZOE over a 90-day period in challenging winter conditions aboard the Ambition cruise ship. Jamie Bartnett of International Cruise Victims described the advancement as a significant step forward in safety, allowing for the quick deployment of smaller rescue vessels once the precise location and time of a fall are known.

Zelim's CEO, Sam Mayall, founded the company following a personal experience that highlighted the long delays in detecting overboard incidents. While traditional methods rely on human observation, modern cruise lines are increasingly employing advanced sensor systems. The ZOE system requires between 12 and 26 cameras for full 360° coverage, depending on the ship's size. The system has now achieved certification, passing international safety standards, including ISO standards, with a demonstrated high detection accuracy and a low false alarm rate, even in difficult conditions.