This reduces the amount of
erroneous pixel classification
that occurs early in the sequence due to the background being unknown.
Figure 1 shows the results of the segmentation using background subtraction.
The black pixels indicate the background, and the white areas
indicate the foreground, i.e. where the bees in the frame are.
Figure 2 shows the same frame with contour and bounding
Contours surrounding foreground and their bounding boxes