Acquisition of traffic information using a video camera with 2D spatio-temporal image transformation technique

Road observation is a very important work for traffic information collection, regularizing traffic, traffic management and urban planning. In this paper, we present a new method for traffic information acquisition from images taken by a camera set over a road, and a 2D spatio-temporal image technique for vehicle detection and description in road observation is discussed. In this research, our objective is not only to detect the vehicles on a road, but also to acquire 3D information about individual vehicles, which is useful for vehicle classification and description. 3D vehicle structures include many horizontal lateral lines. These lines include information for determining the vehicle structure. In the processing, we take these horizontal lines of vehicles as specific features for vehicle detection. The 3D information of the feature lines is obtained from the 2D spatio-temporal image created from the images taken by the camera over the road