This research paper demonstrates the potential of using paint markings for automatic honey bee behavior analysis in the field. A novel dataset for bee re-identification with paint markings is contributed, with nearly perfect recognition in closed settings. The paper also explores the potential for full automation of visit detection and provides preliminary results for future real-time deployment in the field on an edge device. The study highlights the crucial role of honey bees in modern agriculture systems and the need for improved methods to monitor and study these pollinators.
Publication date: 14 Nov 2023
Project Page: Not Provided
Paper: https://arxiv.org/pdf/2311.07407
