The potential of close-range photogrammetry in evaluating the severity of road surface deformations

Abstract

This study evaluates the use of close-range photogrammetry, specifically the Structure-from-Motion (SfM) technique, for assessing road surface deformations. Utilizing a low-cost GoPro Hero 10 action camera, 3D models were created for a 45-meter degraded road section, in order to calculate the volume of road surface defects. Three image collection strategies were compared: only near-nadiral imagery, and near-nadiral plus one or two tracks of oblique images, respectively. Reference data was obtained from Terrestrial Laser Scanning (TLS), in order to assess the level of accuracy for deformation volumes calculated from SfM data. The photogrammetric models, scaled and georeferenced using Ground Control Points (GCPs), allowed for an accurate assessment of deformation volumes (RMSEs of 0.16–0.18 dm3 for the 32 potholes identified in the studied sector) and road surface reconstruction (with average displacements between SfM and TLS point clouds of 0.007–0.13 meters). Deformation volumes extracted from SfM 3D models of the road surface are highly correlated with reference volumes from TLS data, regardless of image collection strategy. However, using at least one track of oblique image collection leads to an accuracy improvement. Our findings confirm that close-range photogrammetry with a low-cost, easily available action camera, is an effective, cost-effective alternative for monitoring road deformations, offering high-resolution 3D models suitable for precise volume and depth measurements. However, the technique does have some limitations, mainly related to the need of GCPs in order to scale the 3D models and the significant amount of manual labour necessary in order to calculate the volume of road surface defects.