Tracking termites: computer vision tools for termitologists

Quantifying movement patterns of organisms creates the possibility of a scientific
and data driven analysis of recurring behavioral events and interactions within an
insect population under study. With the modern technological advances, gathering data
of this kind is made easier in the sense that a larger number of tools and techniques
are at disposal of the interested party.  Here we present a set of these tools created
using digital image processing and computer vision methods, of tracking termites
on experiments video footage made at the Federal University of Viçosa Termitology
Lab. Using the Python’s language OpenCV library, the devised package submits an image
sequence to analysis, identify the termites present at the movie, and tracks these
individuals along time, reporting their position in the frame, the distances between
them and who they are colliding with. The collected information is then presented
to the user in a useful graphical form. For a performance demonstration, a case study
is presented at the end of this paper where we show the movement analysis output
of a termite mold population sample recorded at a Petri plate during one of our
routine experiments.