Cooperative Fish and Wildlife Research Units Program: Missouri
Education, Research and Technical Assistance for Managing Our Natural Resources

Tang, Z., Y. Zhang, Y. Wang, Y. Shang, R. Viegut, E. Webb, A. Raedeke and J. Sartwell. Drones and Machine Learning Integration in Waterfowl Population Surveys. Proceedings of the IEEE International Conference on Tools With Artificial Intelligence


The rapid technological development of drones has led to an increase in capabilities of aerial image collection and analysis for wildlife monitoring. Historically, wildlife abundance estimates were based on manual counts from the ground or from the air using fix-winged aircraft over the regions, both of which are expensive and potentially dangerous methods. However, drones can help easily collect aerial images with a limited budget and increased flexibility, as it creates less disturbance to the wildlife, allowing us to get closer to the target animals and providing clearer images. In this paper, we propose a new, integrated system of drones and machine learning for waterfowl population surveys, which provides a user-friendly interface for data collection and integrates data post-processing using deep learning methods to detect and count waterfowl automatically. Our system has proved to be an efficient and accurate approach of collecting, analyzing, and providing outputs of waterfowl abundance estimates using drones and machine learning.