In order to reduce the lane departure accident caused by driver’s negligence, lane departure warning system (LDWS) has become increasingly prevalent and important. Nowadays, most approaches proposed mainly focus on how to detect lane markings by single-vehicle system. However, due to limited detection range and processing ability, the robustness of single-vehicle systems is poor in complex traffic scenarios. Therefore, this paper proposes a method of lane departure identification based on roadside multi-sensor fusion. The time to lane crossing (TLC) is used to identify the lane departure. Compared with the single-vehicle system, the roadside equipment has stronger robustness in special scenarios. And this approach avoids the calculation resource consumption of single-vehicle system caused by lane detection. The experiment verified that the vehicle detection rate of various scenes exceeded 98%, the lane departure identification rate was 100%.

CICTP