TAIPEI (Taiwan News) — National Cheng Kung University (NCKU) has developed Taiwan's first AI-powered drowning detection system.
On Saturday (Nov. 9), NCKU conducted a live test of the AI-based ocean drowning detection system at Tainan's Anping District Sunset Platform, per CNA. The system is expected to become a valuable tool for lifeguards in monitoring and rescuing swimmers.
Developed by Professor Doong Dong-jiing (董東璟) from NCKU’s Department of Hydraulic and Ocean Engineering, the system utilizes high-precision optical cameras and deep learning algorithms to identify individuals potentially in distress or calling for help near the shore. When the system detects distress signals, such as waving, prolonged submersion, or crossing the designated swimming boundary, it alerts on-site lifeguards or notifies management personnel to initiate rescue protocols.
The system offers five major functions: monitoring rip currents, tracking beach crowd density, recognizing different ocean recreational activities, detecting boundary crossings, and identifying drowning signs.
During Saturday’s demonstration, the system distinguished between various activities such as windsurfing, stand-up paddleboarding, jet skiing, and swimming. When a swimmer crossed the safety boundary, the system marked them with a yellow alert box. A sailing association coach then simulated a drowning scenario, which the AI system detected instantly, highlighting it with an emergency red box.
Doong explained that with hundreds or even thousands of beachgoers, it’s challenging for lifeguards to monitor every second. Once installed, this system can transmit real-time detection results to lifeguards, enhancing their ability to respond to emergencies.
This is the first use of AI technology to improve safety in Taiwan’s ocean recreational areas. Doong said the system has been adapted to Taiwan’s diverse ocean environments and can function under various sea conditions; in addition to ocean settings, the system can be used in swimming pools, streams, ponds, and reservoirs, greatly improving water safety.
NCKU said the team has completed multiple field tests, demonstrating the system’s high accuracy and quick response in detecting recreational activities and drowning risks. The team hopes this system can be implemented nationwide to support government efforts in strengthening ocean rescue capabilities and reducing drowning incidents, showcasing Taiwan’s progress in applying AI technology to public safety.
