In recent years, automated driving technology has improved, and in some regions automated vehicles with practical applications are in operation.

As the practical application of automated driving has many implications, there are high expectations for its practical application in Japan.

However, unlike other countries, Japan has many level crossings in urban areas, so automated vehicles need a system to safely pass through level crossings. In addition, it is necessary to check not only the state of the level crossing but also whether a train is actually approaching, considering the fact that there have been accidents where the level crossing circuit breakers have not been lowered and that there are a small number of level crossings that do not have circuit breakers.

To solve this problem, this study considers a train detection technique using the OpenCV cascade classifier. The practical application of this technology is expected to bring us one step closer to the operation of automated vehicles in Japan.

The application of this technology could also improve traffic congestion and fuel consumption at level crossings, which are problems at level crossings in Japan. Japanese road traffic law requires vehicles to pause in front of level crossings to check whether it is safe to do so. However, all vehicles have to pause in front of level crossings, causing congestion and thereby reducing fuel consumption. Human-driven vehicles need to pause to check for safety, but by eliminating the need for a pause by having an automated vehicle check for safety in a split second, traffic congestion can be reduced and fuel consumption improved.

This study investigated the level crossing in front of Shin Kiryu station, where traffic congestion is actually caused by level crossings.