Pollinators as Data Collectors: Estimating Floral Diversity with Bees and Computer Vision

Frederic Tausch, Jan Wagner, Simon Klaus; Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2023, pp. 643-650

This paper presents a bee-based environment monitoring system that uses pollen color analysis to estimate floral diversity. The study focuses on non-invasively assessing pollinator habitat quality using computer vision technology on honey bee hives. By strategically placing cameras at the beehive entrance, the system captures pollen color samples without disrupting the bees' natural foraging behavior. The collected pollen color data is analyzed using computer vision techniques, including pollen color classification and diversity assessment. The feasibility of the approach is evaluated through comparisons with laboratory analysis results and an appliance for capturing pollen color under ideal conditions. The study also includes the creation of a dataset for further research and advancements in the field of floral diversity estimation. The findings demonstrate the potential of using bees and computer vision technology for monitoring and understanding pollinator habitat quality.