Design and Estimation of the Accuracy Parameters of a Digital Surface Model (DSM) Using a DJI Phantom 4 Pro Drone and Photogrammetry Technique

Abstract

A Digital Surface Model (DSM) represents terrain elevation, incorporating natural and artificial features within an environment. These models are valuable in fields like telecommunications for analyzing electromagnetic wave propagation. However, DSMs are approximations due to errors from acquisition tools, data collection conditions, and processing methods. This study presents a methodology for producing a DSM and calculating its accuracy using a DJI Phantom 4 Pro drone at altitudes of 30 m, 50 m, and 70 m. Images captured were processed with photogrammetry techniques using Pix4Dmapper software. For the respective altitudes, 374, 154, and 80 images were collected at a resolution of 5472×3648 pixels. The photogrammetry technique, employing the Triangulated Irregular Network (TIN) interpolation method, achieved optimal accuracies during the 30 m flight, with planimetric accuracies of 13 cm in the East-West direction and 17 cm in the North-South direction, and distance and height accuracies of 12 cm and 15 cm. This study demonstrates that high planimetric precision and accurate ground structure representation are achievable using drone-based DSM generation.