A traffic avoidance system based on binocular ranging

To automatically detect the front obstacles of smart car, a traffic avoidance system based on binocular ranging is presented in this paper. The whole system consists of two cameras, an image processing system, an intelligent car control host, and a two-drive tracked car. Specifically, two cameras are calibrated to get the distortion parameters, internal parameters, relative rotation matrix and translation matrix parameters. Then the stereo matching algorithm is applied to obtain the disparity map of the scene, and the three-dimensional scene is reconstructed in real time. Finally the distance of front obstacles is obtained, which can be used to achieve obstacle avoidance operation. The experimental results demonstrate that the proposed system can identify the front obstacles and achieve real-time obstacle avoidance effectively.