Automated Counting Wild Birds on UAV Image Using Deep Learning

The monitoring of migratory geese at known stopover sites is crucial to their habitat conservation but usually requires skilled manpower for counting large flocks of waterfowl. We used a UAV to count greater white-fronted geese (Anser albifrons). A single UAV flight could observe the entire lake from an altitude of 100 m above the water surface with little disturbance to the roosting geese. We used a deep learning to automatically count geese in the imagery. The counting accuracy ranged from -7% to -3% in three validation cases compared with manual counts on the UAV image. We conclude that the combination of UAV and deep leaning methods can yield goose counts with an accuracy of ±15%. The results suggest that this approach will be useful for monitoring geese or other waterfowl.