Meeting Abstract
Ecological and behavioral studies and conservation efforts are often complicated by the need to gather data in remote or inaccessible areas. In military missions, unmanned aerial systems (UAS) have been instrumental in providing remote access and persistent presence, including at night. We used a thermal imaging cameraoperated from a UAS to study olive ridley sea turtles (Lepidochelys olivacea) at night during an arribada in Ostional, Costa Rica, in the spring of 2019. During 59 flights and approximately 5 hours of air time spread over two evenings, we repeatedly surveyed a 0.8 km (0.5 mile) transect along the main nesting beach in the Ostional National Wildlife Refuge. We observed the turtles, their tracks, and nests, as well as some known predators on sea turtle eggs, all of which have distinctive signatures in visible and thermal imagery. Hot, subsurface sand provides nests and tracks with a particularly conspicuous thermal signature. Post mission image processing includes photostitching of geotagged thermal imagery and use of machine vision, Haar cascades, deep learning, and convolutional neural networks to identify and count the turtles as well as other animals of interest. The field team consisted of engineering students in a “School of Drones” curriculum at a primarily undergraduate institution, who were presented with a challenge to develop tools for biological study and to advance use of UAS in science missions by “owning the night” and engaging with scientists and local partners.
