Digital Terrain Modeling across Varied Terrains: UAV Photogrammetry vs. Total Station Surveys

Digital terrain models (DTMs) are fundamental to engineering surveying and mapping.Unmanned aerial vehicle (UAV) photogrammetry has emerged as an efficient, low-cost alternative to conventional ground surveys, delivering very high-resolution 3D surface data with minimal field labor.Under ideal conditions, wellplanned UAV surveys can achieve centimeter-level accuracy comparable to traditional methods.However, their performance can degrade on steep or vegetated terrain.In this study, we compare DTMs generated by a DJI Mavic Air 2 UAV (processed in Pix4D) against those from a Trimble total station, across three contiguous sites at Kathmandu University: a flat plain, an undulating hill, and a semi-urban settlement area.The UAV-derived DTM achieved the highest accuracy (lowest RMSE) on the flat terrain and exhibited a significant elevation bias on the steep hillside.Specifically, the UAV DTM closely matched the total-station DTM on the plain and settlement sites (near-zero mean difference), whereas the hilly site showed a large systematic offset due to georeferencing issues.These results indicate that UAV photogrammetry is highly effective for flat, open areas but may require additional ground control or hybrid methods in complex environments.The findings provide practical guidance on when drone-based surveys can reliably replace or augment traditional surveying.