Following Vehicle Detection Using Multiple Cameras

In recent days, various in-vehicle camera systems have been proposed for detecting following vehicles. In this paper, we propose a vehicle detection method based on feature points. First, the feature points are detected from the image obtained from the camera and the optical flow is calculated by tracking the feature points across the image sequence. Then the feature points are assumed to be coplanar and to model the front of the following vehicle. This plane's movement can be defined by an affine-transformation, and the optical flow of the plane is used to determine whether a section of the image is in the background or is part of the following vehicle. By extending previous detection systems based on edge models to include feature point tracking, our method is able to produce improved detection of following vehicles. To verify the accuracy of our method, we have performed experiments in a variety of different road conditions and situations. As a result, the proposed system is expected to be useful for reducing traffic accidents caused by rear- and/or side-located vehicles.