Dataset from the results of an experiment to determine how three controllable factors, flight altitude, camera angle, and time of day, affect human identification and counts of animals from drone images to inform best practices to survey animal communities with drones. We used a drone (unoccupied aircraft system, or UAS) to survey known numbers of eight animal decoy species, representing a range of body sizes and colors, at four GSD (ground sampling distance) values (0.35, 0.70, 1.06, 1.41 cm/pixel) representing equivalent flight altitudes (15.2, 30.5, 45.7, 61.0 m) at two camera angles (45° and 90°) and across a range of times of day (morning to late afternoon). Expert human observers identified and counted animals in drone images to determine how the three controllable factors affected accuracy and precision. Observer precision was high and unaffected by tested factors. However, results for observer accuracy revealed an interaction among all three controllable factors. Increasing flight altitude resulted in decreased accuracy in animal counts overall; however, accuracy was best at midday compared to morning and afternoon hours, when decoy and structure shadows were present or more pronounced. Surprisingly, the 45° camera enhanced accuracy compared to 90°, but only when animals were most difficult to identify and count, such as at higher flight altitudes or during the early morning and late afternoon. We provide recommendations based on our results to design future surveys to improve human accuracy in identifying and counting animals from drone images for monitoring animal populations and communities.
College of Forest Resources
Department of Wildlife, Fisheries and Aquaculture
animal community, AGL, animal survey, best practices, camera angle, count bias, flight altitude, RPAS, shadows, UAS, UAV
Ecology and Evolutionary Biology
Jones, Landon R.; Elmore, Jared A.; Krishnan, B. S.; Samiappan, Sathishkumar; Evans, Kristine O.; Pfeiffer, Morgan B.; Blackwell, Bradley F.; and Iglay, Raymond B., "Dataset for Controllable factors affecting accuracy and precision of human identification of animals from drone imagery" (2023). College of Forest Resources Publications and Scholarship. 24.