(Pace, Jenessa; Ence, Elodie; Kunz, Allison; Stevens, Rebecca; Kunz, Amber; Cooper, Izaak; Nelson, Alicia; Asplund, Alyssa; Stock, Julianna) (Brigham Young University)
Faculty Advisor: Williams, Gustavious (Ira A. Fulton College of Engineering and Technology, Civil and Environmental Engineering)
Photogrammetry using Unmanned Aerial Vehicles (UAV) has become more relevant to water resources issues. The models generated can be highly accurate and detailed. The use of ground truth is an integral part of ensuring the accuracy of such models. Ground truth based on GPS-obtained data of visual targets gathered in the field prior to UAV data collection are used to improve post processing accuracy. Proper spacing of ground truth is primarily dependent on the flight path of the UAV during data collection, flight height, and image resolution.
One developing task is creating models of full-pool bathymetric maps for reservoirs - using UAVs for above the waterline and sonar for below. Due to their erratic shapes, capturing the shorelines of reservoirs requires erratic flight paths. Planning ground truth spacing requires special care.
To determine proper ground truth spacing we completed a field survey at Starvation Reservoir, near Duchesne, Utah. We selected a two mile stretch along the north side of the reservoir to test various ground truth spacings. We flew the test stretch with two flights, one taking nadir photos and one with the photos taken at an angle of about 30 degrees, more normal to the shoreline. The primary placement of ground truth points was approximately 0.25 miles apart. After the data was recorded, we created models using Metashape software using different ground truth spacing, using spacings of 0.25, 0.5 and 1 mile. We used Cloud Compare software to determine the error between each model. We assumed that the model generated using 0.25 points/mile was "truth". We used this preliminary information to determine that for bathymetry maps, a ground truth spacing of 1 mile/point was acceptable, we are continuing our research to refine this finding.