This work performs visual localisation using synthesised local orthographic imagery. We exploit state of the art stereo visual odometry (VO) on our survey vehicle to generate high precision synthetic orthographic images of the road surface as would be seen from overhead. The fidelity and detail of these images far exceeds that of aerial photographs. When traversing the same route at a later time, the vehicle is localised against the survey vehicle’s trajectory by maximising the mutual information between the synthetic and live image streams,  explicitly leveraging dense appearance rather than a sparse set of point features, allowing for centimeter precision.

 fig5


  • [PDF] A. Napier and P. Newman, “Generation and Exploitation of Local Orthographic Imagery for Road Vehicle Localisation,” in Proc. IEEE Intelligent Vehicles Symposium (IV), Madrid, Spain, 2012.
    [Bibtex]

    @inproceedings{AshleyIV2012,
    Address = {Madrid, Spain},
    Author = {Ashley Napier and Paul Newman},
    Booktitle = {Proc. IEEE Intelligent Vehicles Symposium (IV)},
    Date-Added = {2012-03-28 09:00:00 +0000},
    Keywords = {Road Navigation With Overhead Priors},
    Month = {June},
    Owner = {Ashley},
    Pdf = {http://www.robots.ox.ac.uk/~mobile/Papers/2012IV_napier.pdf},
    Timestamp = {2012.03.28},
    Title = {Generation and Exploitation of Local Orthographic Imagery for Road Vehicle Localisation},
    Year = {2012}}