Automatic Detection and Mapping of Highway Guardrails from Mobile Lidar Point Clouds.

Accurate and rapid detection and mapping of guardrails play an essential role in road maintenance, asset management, and safety audits. This paper proposes an automated approach for guardrail detection and mapping in rural highway environments from mobile laser scanning point clouds. First, the scanning vehicle trajectory is extracted, and a buffer zone is constructed to remove points far from the roadside. Then multiple geometric filters and statistical outlier removal are applied to find the candidate guardrail points. The verticality of each point is calculated to further refine the detection. Next, a density-based clustering algorithm is used to group the remaining points into clusters. Finally, the detected guardrails are plotted on Google Earth to create an inventory map. Experiments on three rural highways demonstrate that the proposed algorithm can detect and map guardrails automatically with high accuracy.