Tracking Bees Using Imaging Analysis
Oral Defence Date:
Profs. Bill Hsu & Ilmi Yoon
Bees play an important ecological and agricultural role as pollinators. Bee colonies are large, well-organized systems; monitoring the health of colonies is likewise important. Video footage is often used to monitor bee activity. However, one of the difficulties in tracking bee activity over an extended period is the monotonous nature of the task. Human attention will naturally wander during the duration of a lengthy video. Automatic software based tracking using computer vision techniques may be useful for monitoring bee activity in extended collections of video footage. Image analysis techniques can even extract useful information from video footage, greatly automating the normally time-consuming process of scanning footage for objects and events of interest. The software can annotate the footage automatically for later review. This paper introduces the BeeTracker program, a cross-platform application that uses image analysis techniques to track bee activity. We focus on two types of bee activity that may be correlated with the health of a bee colony: hive entries and exits, and waggle dances. BeeTracker is easy to use, and able to accurately identify these two types of bee activity; hence, it may be a useful tool for monitoring the health of colonies.
Java, Processing, computer vision, image analysis, blob detection, bee, Eclipse.