A video-based approach for stationary platform supervision

The paper presents a stationary system developed for automatic supervision of railway station platforms, especially under conditions of automated train operation. It uses cameras mounted at the platform and supervision computers analyzing the camera data to detect persons or obstacles in different platform areas. The paper starts with a review of the state of the art and derives the system task and approach. Then, an overview of the system is given with its hardware and software technology, image processing, and operational surroundings. Practical test results and an evaluation conclude the paper.