Discrepancy Analysis between Close-Range Photogrammetry and Terrestrial LiDAR

Abstract

This work presents a comparison of spatial measurements performed on 3D virtual models of civil structures against actual field measurements completed on the related real objects. The virtual models were generated by two different techniques. In one instance, the 3D models were produced via close-range photogrammetry (CRP). Such models are based on numerous still-frame 2D photographs, post-processed with commercially available photogrammetric software. In the second approach, 3D point-cloud models were generated via terrestrial light detection and ranging (LiDAR), laser scanning. For comparison purposes, two case studies were conducted. The first involved a single story structure at the campus of Georgia Southern University. The second case was a multi-level Maya ruin at Dzibilchaltun, Merida, Mexico. For accuracy comparisons, various benchmarks were established around the structures following accurate closed-traverse procedures. The benchmarks served as standard georeferenced points. Several physical target points were then marked on the exterior walls of the structures. They are referred here as reference wall points (RWPs). The actual RWPs were then measured with typical laser-based total-station instruments. After photographs were taken and laser scanning of the structures were completed, the coordinates of the RWPs points were also determined from the respective virtual models. The virtual coordinates were then compared against the ones obtained with the total-station instruments. Coordinates and distances from each procedure were compared to determine discrepancies. Results of this study demonstrate that close-range photogrammetry can provide accurate enough information to be employed as an alternative 3D measuring and modeling technique for surveying and civil engineering applications involving structures and areas within the size range covered in the presented case studies. Additionally, several recommendations on the use of CRP and laser scanners are presented.