The dynamics of individual bacteria underlies the manifestation of complex multicellular behaviours such as biofilm development and colony expansion. High resolution movies of expanding bacterial colonies reveal intriguing patterns of cell motions. A quantitative understanding of the observed behaviour in relation to the bacteria's own motile apparatus and to hydrodynamic forces requires that bacteria be identified and tracked over time. This represents a demanding undertaking as their size is close to the diffraction limit; they are very close to each other; and a typical image may contain over a thousand cells. Here, we describe the approach that we have developed to segment individual bacteria and track them in high resolution phase contrast microscopy movies. We report that over 99% of non-overlapping bacteria could be segmented correctly using mathematical morphology, and we present preliminary results that exploit this new capability.