Hitachi Rail has announced the establishment of IARO (Artificial Intelligence for Obstacle Recognition) – a new safety system designed to detect obstacles at level crossings.
The new solution, which has now been validated for use on the Adif rail network, consists of a CCTV camera and processing unit that analyses images in real-time.
Once analysed, the system warns those controlling the train of the presence of obstacles within the level crossing area before it arrives within the zone.
Utilising deep learning technology, the IARO system has been trained on hundreds of hours of video and images of people, cars, motorcycles and other vehicles from various angles, allowing for the detection and classification of obstacles regardless of their position, shape or size.
The system is capable of automatically classifying obstacles and discarding those that pose no danger, such as static objects already present within the area.
With the certification of Adif; IARO is classified to have officially met the established requirements in place for Obstacle Detection Systems for Level Crossings. The system recently passed a technical phase validated by an independent company, as well as another phase of both testing and certification in laboratories approved by Adif before passing field tests at the Pozaldez level crossing in Valladolid.
