Method for Detecting Cars Cutting in to Change Lanes by Using Image Frames Played in Reverse

摘要

In this paper, we suggest a method for tracking cars that violate traffic laws restricting cut in lane changes. We also present experimental results, which show that the tracking method can be implemented in unmanned instruments for detecting violating car by using the image frames played in reverse. Two kinds of camera were installed as a set at the Yang-Jae IC in Korea: A recognition-camera (R_c), and A tracking-camera (T_c). The R_c reads a plate number and the T_c tracks cars in a region of interest (ROI). The T_c determines whether or not the cars violate the law by analyzing image frames played in reverse. A plate number recognition algorithm was provided by KNU DILAB, and the KLT algorithm was used for tracking cars in the ROI. Our experimental findings show that the proposed method can be applied to unmanned systems for cars that illegally cut in to change in lanes.