MEASURING RIVER GEOMORPHIC CHANGE USING A SMALL UNMANNED AIRCRAFT SYSTEM

Understanding the evolution of New England river channels informs emergency management procedures, municipal planning, and stream restoration efforts. Recent technological advances hold the promise of providing quicker, cheaper, and higher resolution methods for quantifying river geomorphic change through serial comparison of digital elevation models (DEMs). This study uses digital photographs taken from a consumer-grade small unmanned aircraft system (sUAS) to construct DEMs via the software-based Structure from Motion (SfM) photogrammetric technique. We use this “sUAS SfM” method to survey the topography of study reaches on several rivers in Vermont and New Hampshire. Preliminary results show sUAS SfM surveys consistently produce DEMs and orthophotographs with ~8 cm pixel resolution. Surveys using a laser total station (TPS; with relative accuracy on the order of a few mm) with integrated global positioning system (GPS) were conducted with sUAS SfM surveys in order to georeference and assess the accuracy of the resulting sUAS SfM-based DEM. Although sUAS SfM surveys appear to faithfully reproduce topographic variation, one-to-one point comparisons show sUAS SfM surveys measure elevations 0.23 ±0.21 meters (m) higher than TPS equivalents. In order to evaluate GPS-introduced error of TPS measurements, we surveyed ground control points (GCPs) used for georeferencing from multiple TPS stations. Comparisons of absolute positions of 18 point pairs indicate longitude, latitude, and elevation (X, Y, and Z) TPS/GPS root mean squared (RMSE) error of 0.03, 0.03, and 0.04 m respectively. The magnitude of TPS/GPS error is therefore much less than sUAS SfM error and it appears sUAS SfM measurement error may arise from the SfM processing algorithm, an observation consistent with similar studies. However, changes in image acquisition height, image processing, or other modification to the sUAS SfM workflow may mitigate sUAS SfM topographic measurement error. A major goal of this study is to understand how error in sUAS SfM elevation data might be overcome. Further work will also focus on leveraging the rapid survey ability of sUAS SfM to detect geomorphic change in river systems in response to or recovery from flood events.