A Monocular Ranging Algorithm for Detecting Illegal Vehicle Jumping

In order to detect the illegal vehicle jumping in the crowded traffic flow by the car DVR, an algorithm for real-time ranging of the car in front using license plate characteristics is proposed with the monocular vision technology. Firstly, the geometric relationship model of the camera is studied, and then a kind of real-time ranging algorithm with the monocular vision is put forward combining the license plate characteristics. Based on the study of the geometric relationship between the pixel size of the marked area and the plane of the camera, put forward to use the pitch angle of the camera to correct the size of the pixel in the license plate recognition region, and to further improve the accuracy of the ranging. The experiment results shows that the improved method, being accurate and robust, is able to meet the real-time requirements of distance ranging at both static and dynamic tests.