State DOT’s manage millions of valuable roadway assets, spread across hundreds of millions of acres. Inventorying assets in the field is costly and time-consuming, especially when accurate geolocation is included. Remote sensing may offer alternatives to field inventory, especially if the process can be automated through imagery classification. Unfortunately, many roadway assets—signage, signals, luminaires and guardrails, for example—present spectral profiles that are difficult or impossible to identify through automated classification of 2-dimensional photo imagery. Likewise, aerial lidar only captures an asset profile as seen from above. Mobile lidar captured from specially-equipped (often purpose-built) vehicles, on the other hand, captures rich, 3D representations of roadway assets, which may present new opportunities for identifying assets and their locations.
In this series of videos, I introduce a novel approach to extracting roadway assets from mobile lidar point clouds using the workflow pictured. This video will cover the first step in the workflow.