Automatic Road Feature Extraction from State-Owned Mobile LiDAR Data

Accurate, up-to-date road feature data is important for asset management, multi-modal traffic planning, and data-driven traffic safety analysis. Collecting such road features manually is tedious and leaves room for human error. With increased adoption of LiDAR technologies to collect roadway data from vehicles, automatic methods to extract road features is more feasible.

This project used a subset of mobile LiDAR data to develop an algorithm to automatically extract guardrails, lane numbers, and lane widths. The algorithm was packaged as an ArcGIS Tool where LiDAR cloud point data can be inputted. The output consists of feature classes of the guardrails geolocations and the number of lanes and lane widths.