Impacts of Abrupt Terrain Changes and Grass Cover on Vertical Accuracy of UAS-SfM Derived Elevation Models

Recent years have seen an increase in the use of unmanned aircraft systems (UAS; drones) for capturing aerial imagery and processing the images into digital terrain (DTM) and surface models (DSM) using structure from motion (SfM) techniques. However, many questions remain related to the accuracy of these methods across varied terrain and land covers. This study addresses two of these questions by investigating the accuracy of a UAS-derived DSM in (1) an area of abrupt terrain changes with (2) various types of grass cover. Aerial images were captured over a mesa in Gloss Mountain State Park, near Fairview, Oklahoma. The images were processed with SfM to generate a DSM. Field plots were located to capture various measurements of grasses including height, leaf area index, and species. The accuracy of the DSM was assessed to determine how the severe terrain changes and different grassland-vegetation covers influenced findings. Results show that flight direction and varying terrain-to-platform distance may increase errors in regions of abrupt terrain. Meanwhile, the diverse grass cover characteristics were able to explain 54 percent of the variance in DSM accuracy through least squares regression.