抄録
Slow-speed-notifying signals are essential safety-related components of railway signal systems. However, if a backup system such as ATS is not installed, the detection of such signals becomes prone to errors as it depends exclusively on the driver's vision. To address this problem, we are developing a system based on an image processing approach that uses a pattern-matching algorithm. This system is able to recognize slow-speed-notifying signals automatically from sequences of images showing the area ahead of the train, then triggers alarms to assist the driver. This paper describes the details of the approach and outlines the results of tests performed with a prototype program.
