Behaviour 2019
Advanced image composite technique allows efficient collection of ant traffic data
Woncheol Song1, Keesan Lee1, Sang-im Lee2, Piotr G. Jablonski1,3. 1Laboratory of Behavioral Ecology and Evolution, Seoul National University, Seoul, , Korea; 2Laboratory of Integrative Animal Ecology, DGIST, Daegu, , Korea; 3Museum and Institute of Zoology, Polish Academy of Sciences, Warsaw, , Poland

Composite photography has been used for a long time by behavioral biologists who need to track animal movement and traffic, with ants being salient examples. However, the advent of modern graphics hardware and machine-learning algorithms almost obsoleted the composite in current literature, in preference to direct tracking of the trajectories of object centroids detected by the computer. However, we found that the combination of these old and new techniques result in surprisingly affordable and efficient solutions for monitoring small and slow-moving animals such as ants. With modern image processing algorithms such as scale-invariant feature transform (SIFT) and sub-pixel image localization, our method provides researchers with tools for overcoming common field limitations of composite imaging methods, including variable light conditions, obscuring mobile objects, drifting frames, and lack of better resolution. Successful use cases in an interspecific competition study and a traffic analysis are demonstrated, with particular emphasis on tracking small species of Monomorium and Lasius from distances almost twice farther than what can be achieved by unprocessed filming.