Automated Beehive Surveillance Using Computer Vision

The number of honey bees entering and leaving a hive throughout the day is an important metric for beekeepers monitoring the health of their hives. Activity at the hive entrance depends on many factors such as weather, colony health, and presence of an adequate food supply. Some commercial systems that utilize infrared sensor gates at the hive’s entrance exist for accurately measuring honeybee traffic. In this thesis, a solution that is based on visual information obtained through videos taken in front of the hives is explored. This thesis provides details on the design and implementation of a beehive surveillance system, describing the algorithms and techniques used. The system operates in a realistic outdoor apiary environment, providing constant surveillance in real-time. Results obtained using different algorithms within the system are compared and methods for evaluating performance are discussed.