Video Detection in Road Tunnels

Photo and digital image of a truck in a tunnel Copyright: © ISAC

The institute has extensive experience in the use of video detection systems in road tunnels. In several research projects, different video detection systems were qualitatively investigated, which led to recommendations for their use in safety systems of road tunnels. Different application concepts and potentials of the video analysis were analyzed.

The self-developed video detection system isac-DVA is used for PC-based digital video processing (Image Processing) for traffic monitoring and incident detection. It deals with aspects of accident management such as traffic control systems and safety devices in tunnels.

Our software solution "isacVITO"

isacVITO is a real-time computer vision software solution that detects and tracks vehicles in individual sections of the camera. The detection and classification is realized by implementation of a Chamfer matching algorithm. This algorithm compares and evaluates the edge information in the image with predicted geometries of two vehicle classes. For the second step of image processing - vehicle tracking - a tracking based on particle filter algorithms is currently available. The implemented method shows good results for tracking over longer distances of over 100 m. The AUTUKAR project extends the idea of locating individual vehicles in consecutive sections. This is achieved through sensor and data fusion, allowing information such as position, orientation, speed and characteristics of the individual vehicles to be detected. This recorded, highly accurate and distance-related vehicle composition provides the tunnel operators with a tool for determining and assessing the position as well as for the early detection of safety-relevant incidents and follow-up events.