The Relevance of Compound Events in Bee Traffic Monitoring
Authors/Creators
Description
Bees play a crucial role in human agriculture, making the accurate characterization of their behavior essential. In the field of automatic bee traffic tracking, individual monitoring has been prevalent in the past two decades to collect detailed data on foraging behavior. To support analysis, these data must be classified into entry and exit events, and assembled into flight trajectories. However, existing classification methods largely consider only simple exits and entrances, which ignore the existence of U-turn or guarding movements which represent a non-trivial fraction of bee activity at the hive entrance. Identifying these movements
as compound events is necessary to accurately track all movements over time while avoiding classifications which may produce misleading movement records. In this presentation, we analyze existing event classification methods to evaluate their accuracy for two different types of movements: simple and compound. We then propose extended event classification methods that explicitly account for compound events, enabling more accurate reconstruction of bee traffic dynamics.
Files
proc2026.pdf
Additional details
Funding
- U.S. National Science Foundation
- CISE-MSI: DP: SaTC: CyIndiBee - CyberInfrastructure for video analysis of individual bee behavior 2318597