Postdoctoral Research Associate University of Arizona
Disclosure: Disclosure information not submitted.
The use of photogrammetry via unmanned aerial systems, or UAS, is becoming increasingly popular in agriculture. It is a process by which researchers and producers can greatly increase the efficiency of their programs with only minimal expertise. Nonetheless, scientists and producers that hope to determine physical characteristics of their crops must implement some method to georeference the imagery that they use. Ground control points, or GCPs, have been widely used for aerial photogrammetry, but few studies have investigated how to optimize their use. Additionally, although real-time kinematic (RTK) GPS-enabled UAS are beginning to become more affordable for many researchers, studies have not yet been conducted to determine whether RTK is a viable alternative to the traditional usage of GCPs. In this study, we investigated the effect of varying numbers and placements of GCPs and determined that using multiple GCPs improves accuracy of projects without RTK GPS immensely, but accuracy peaks when using four GCPs (2.9cm) when compared to using an entire grid of 63 GCPs (3.0cm). In addition, UAS with and without RTK capability were compared and it was determined that RTK UAS showed greatly decreased error (6.5cm) than UAS without that capability (1.77m) when attempting photogrammetry without the use of GCPs. The effect of GCPs upon flights conducted with RTK UAS was minimal, with only cm-level improvement of accuracy. In contrast, GCPs were much more crucial when analyzing imagery collected with a non-RTK UAS, with m-level improvements of accuracy. These results indicate i) researchers should use the optimal number of GCPs, four in this case, as little benefit is gained from placing more, and ii) the use of UAS with RTK capability may supplant the need for GCPs entirely, indicating that these new UAS platforms will be invaluable to researchers moving forward.