OKI and Marubeni Network Solutions will participate in an introduction test of AI-based level crossing abnormality detection conducted by Nankai Electric Railway.
In this test, ”Level Crossing AI Detection System” jointly developed by OKI and Marubeni Network will be used to perform real-time AI image processing on footage from existing surveillance cameras at level crossings to detect people and bicycles stranded at the level crossing.

When a stagnation is detected, an interlocking signal emitter notifies the driver of an approaching train of the abnormality, helping to prevent level crossing accidents. After verification through this test, Nankai Electric Railway aims to fully introduce the system from fiscal 2024 onwards.

 “Level Crossing Accumulation AI Detection System” uses skeletal detection technology and AI edge technology to process images from cameras installed at level crossings with high precision and in real time on the spot. This is a system that detects people and bicycles staying inside. By linking with local special signal emitters, it provides a one-stop service that detects stagnation, sends a warning to a command center, and even displays a stop signal to drivers, supporting the prevention of railroad crossing accidents.

In this experiment, by using surveillance cameras installed to monitor pedestrian railroad crossings, there was no need to install new cameras, making it possible to easily and inexpensively introduce the system.