Automatic detection of honeybees in a hive

Abstract:

The complex social structure of the honey bee hive has been the subject of inquirysince the dawn of science. Studying bee interaction patterns could not only advancesociology but find applications in epidemiology as well. Data on bee society remainsscarce to this day as no study has managed to comprehensively catalogue allinteractions among bees within a single hive. This work aims at developingmethodologies for fully automatic tracking of bees and their interactions in infraredvideo footage.

H.264 video encoding was investigated as a means of reducing digital video storagerequirements. It has been shown that two orders of magnitude compression ratiosare attainable while preserving almost all information relevant to tracking.

The video images contained bees with custom tags mounted on their thoraxeswalking on a hive frame. The hive cells have strong features that impede beedetection. Various means of background removal were studied, with the median overone hour found to be the most effective for both bee limb and tag detection. K-meansclustering of local textures shows promise as an edge filtering stage for limbdetection.

Several tag detection systems were tested: a Laplacian of Gaussian local maxima basedsystem, the same improved with either support vector machines or multilayerperceptrons, and the Viola-Jones object detection framework. In particular, this workincludes a comprehensive description of the Viola-Jones boosted cascade with a levelof detail not currently found in literature. The Viola-Jones system proved tooutperform all others in terms of accuracy. All systems have been found to run inreal-time on year 2013 consumer grade computing hardware. A two orders ofmagnitude file size reduction was not found to noticeably reduce the accuracy of anytested system.