Reconstruction of Three-Dimensional Trajectories of Honeybees Flying in High-Density Aerial Environments

So far, relatively few studies have investigated the flight of insects moving in experimental conditions approximating natural scenes. This is mainly because of the technical difficulties involved in detecting and tracking individual insects in three dimensions in outdoor scenarios, as well as matching views of corresponding insects in stereo images for trajectory reconstruction (the so-called Correspondence Problem). In this study, we describe the methods we have developed to track and unambiguously reconstruct the trajectories and body orientations of a large number of bees flying in close proximity in a bee cloud, using just two high-speed video cameras configured as a stereo pair in a semi-outdoor setting. Using these methods, two separate bee clouds were filmed and the data were analysed to reconstruct the three-dimensional trajectories, including the head and tail positions, of a total of 382 bees. This dataset should enable future analysis of the movement characteristics of bees flying in a dense environment, as well as uncover potential strategies for mid-air collision avoidance.