Successive Observations of Trajectories of Vehicles with Plural Video Cameras

In this study, a system was developed for successive observations of trajectories of vehicles with images of plural video cameras installed adjacently along roads through the use of techniques such as image processing and geometry transformation. This system can more highly equilibrate the observation precision and the observation range than the use of images of a single video camera. In addition, it can exclude the influence of occlusion where a vehicle is shadowed by another in video images. Furthermore, an observation equation was derived in combination of equations of triangles that approximate road surfaces with equations of a 3D projective transform. The measurement accuracy of this method was evaluated in comparison with results of a precision survey of trajectories of a 1/22 scale vehicle model