Modeling Road Centerlines and Predicting Lengths in 3‐D Using LIDAR Point Cloud and Planimetric Road Centerline Data

Abstract

Abstract: Transportation is one of a few engineering domains that work with linear objects—roads. Accurate road length information is critical to numerous transportation applications. Road lengths can be obtained via technologies such as ground surveying, global positioning systems (GPS), and Distance Measurement Instruments (DMI). But using these methods for data collection and length determination is time‐consuming, labor intensive, and costly. The purpose of this study was to assess the accuracy and feasibility of an alternative. This article reports on a study that provides an alternative to obtaining road centerline lengths by measurement; instead it proposes using geographic information systems (GIS) and light detection and ranging (LIDAR) point cloud data. In this study, a three‐dimensional (3‐D) vector model based on linear referencing systems (LRS) concepts was developed to represent road centerlines in a 3‐D space and to predict their 3‐D lengths. A snapping approach and an interpolation approach to obtain 3‐D points along lines when working with LIDAR point clouds were proposed and discussed. Quality control measures were initiated to validate the approach. The accuracy of the predicted 3‐D distances was evaluated via a case study by comparing them to distances measured by DMI. The results were also compared to road lengths obtained by draping planimetric road centerlines on digital elevations models (DEMs) constructed from LIDAR points. The effects of the average density of 3‐D points on the accuracy of the predicted distances were evaluated. This study concluded that the proposed 3‐D approach using LIDAR data was efficient in obtaining 3‐D road lengths with an accuracy that was satisfactory for most transportation applications.