Meeting Abstract
Honeybees are well known for the highly complex and organized roles they play in colonies of thousands of individuals, yet they have not yet been studied on a large scale to discover the effects of the division of labor on the hive as a whole. We recently selected a colony of A. mellifera carnica at Namik Kemal Üniversitesi in Degirmentalti, Turkey as our study site to investigate the automation of the analysis of multiple complex social behaviors on an individual and community scale in honeybees. This was done by creating a long-term recording network that will be able to track movements of multiple generations of bees for a year long period. Previous studies have used tagging when tracking bees, but have not been able to distinguish between fanning and pollen collecting. We highlighted these behaviors to understand how they impact the circadian rhythm of a single bee, and discover at what time of day they leave the hive to complete these actions. Current monitoring systems like RFID tracking are unsuitable because they can only track a single bee at a time, but by using a modified version of BEETag software, we will be able to track multiple generations of groups of ~500 bees for their entire lifespan. This will be completed through a camera system that will give us 24/7 video of the bees entering and leaving the hive. Current data mining algorithms are not sufficient for video analysis of this complex behavior, which is why we aim to develop a semi-supervised, learning algorithm that can quantify these usually solely manually annotated behaviors. This study design is unique for the year long recording it will facilitate, and discovering which individuals are responsible for nectar/pollen collecting and fanning all in their natural habitat.
