CCTV has been installed in large roadway tunnels in Japan for monitoring of running vehicle and traffic accident. If various sensing function concerning traffic management concentrated in CCTV, both the initial cost and maintenance expenses can be kept to a minimum and the burden of operators can be alleviated. Therefore we reports the Multifunctional Image Processing System concentrated automatic detection of abnormal vehicle behavior, falling objects, and smolder fire, luminance measurement result necessary for lighting control, measurement result of transmittivity necessary for ventilation control, and measurement of various traffic-flow parameters in a traffic control, as part of a disaster prevention system for urban tunnels. With respect to detection of abnormal vehicle behavior, changes in the density of input images are used to extract characteristics of the vehicle, and a target for tracking is identified. The target is tracked based on sequentially input images, and by determining the vehicle behavior, such phenomena as reduced speeds, stoppage, and swerving can be determined. Here traffic flow, average speed, and various other traffic control parameters can be measured simultaneously.
Samenvatting