This paper presents an on-the-fly procedure for obtaining extrinsic camera parameters of onboard vehicular cameras, as well as augmented reality applications for driver assistance systems and road inspection. The proposed approach employs a lane detection algorithm to extract the lane boundaries of the road, and estimates the distance between adjacent lane markings based on the speed of the vehicle. Then, a rectangular portion of the road is used to obtain the desired extrinsic camera parameters. With the complete set of camera parameters, virtual objects can be coherently inserted into the video sequence captured by the camera, so that synthetic traffic signs may be added to increase safety. Experimental results illustrate the performance of the proposed camera calibration systems, as well as possible applications involving augmented reality.

Expert Systems with Applications