Ground-based aerial photography to support traffic flow extraction

Road traffic has been steadily increasing worldwide for several decades. With the rising number of vehicles using the transportation network, the effectiveness of traffic management is becoming more crucial, since new road construction does not keep up with the growth of the traffic volume. As a key for effective traffic management, monitoring traffic flow is of growing interest. Recently, new imaging-based systems have been introduced for traffic flow monitoring. These pole- or bridge-mounted cameras show encouraging performance, and the trend of switching toward imaging technologies is expected to increase. Vehicle (feature) extraction from the imagery is a computer vision task. Due to the oblique viewing angles of the pole-mounted cameras, the imagery is significantly distorted, and therefore to support the feature extraction process, it is important to rectify the imagery. This paper proposes a practical approach for rectifying the oblique imagery to the plane of the road surface. The method does not require an available digital elevation model for the rectification, it only requires measured distances from the camera to at least three points on the road surface well distributed in the image area. The paper provides a detailed description of the proposed method and an initial performance analysis.