Joint Beam Guardrail Detection and Tracking by Lidar for Real-Time Applications

This paper presents a robust and real-time guardrail detection and tracking algorithm, based on the three-dimensional information obtained from lidar sensor. Height feature and corner feature are used to extract guardrail feature points out in guardrail detection process, while a cluster method is used to filter the useless guardrail points. Furthermore, an adaptive region of interest (AROI) is proposed to improve the efficiency and accuracy of the detection algorithm. To determine the boundary of AROI, an improved Kalman filter algorithm is proposed to track the feature point set on both sides of the road at the same time in each frame. The whole system works in real-time and achieves detection accuracy above 96% on a test database.