ASSESSING THE GEOSPATIAL ACCURACY OF AERIAL IMAGERY COLLECTED WITH VARIOUS UAS PLATFORMS

Recent development of small unmanned aircraft systems (UAS) provides a relatively low-cost solution to collect aerial imagery with very high spatial and temporal resolutions. The geospatial accuracy of collected data can range from a few centimeters to several meters, and the use of ground control points (GCPs) is recommended to correct for large geospatial errors. However, whether or not GCPs are used, the true geospatial accuracy of collected UAS data remains unknown. The objective of this study was to measure and compare the geospatial accuracy of images obtained with various UAS platforms at two flight altitudes. Aerial imagery was collected using four platforms equipped with different RGB cameras: Phantom 4, eBee Ag, eBee Plus, and Trimble UX5. All platforms were equipped with manufacturer GPS receivers, and RTK was activated on the eBee Plus. Each platform was flown at 75 and 120 m altitudes, and the experiment was replicated three times. Results demonstrated that using GCPs during data processing improved the horizontal and vertical accuracies of the Phantom 4, eBee Ag, and Trimble UX, decreased the between flight variability, and accounted for the negative effect of flight altitude. On the other hand, the RTK technology used with the eBee Plus resulted in images with very high geospatial accuracy with or without GCPs. Using GCPs during data processing or RTK technology at the time of flight provided aerial imagery with horizontal accuracies of 1.5 to 10 cm and vertical accuracies of 0.0 to 0.4 m. These results are within an acceptable range for data utilization, unlike the horizontal and vertical accuracies obtained without GCPs or RTK, which ranged from 32 to 441 cm and from 1 to 126 m, respectively. Results from this study quantify the geospatial accuracy of UAS imagery and provide a better understanding of the relationships between the accuracy of the GPS receivers in UAS, flight altitude, and horizontal and vertical accuracies of collected images.