Scripts to calibrate a conversion between image and real-world coordinates.
Context: A mobile robot operating in the plane, viewing a set of targets (pucks).
Given the position of a puck in the image (xi, yi), we would like to obtain the
corresponding real-world coordinates (xr, yr). The approach taken here is to
lay out a set of targets in a grid, capture an image, then get the user to
manually select these targets by running correspondences_from_picks.py.
Then run interpolator.py to interpolate between the selected points. The
result is interpolated_correspondences.csv which can then be used as a
lookup-table to convert from (xi, yi) to (xr, yr).
This approach was first described in the following paper:
Vardy, A., Vorobyev, G. & Banzhaf, W. Cache consensus: rapid object sorting by a robotic swarm. Swarm Intell 8, 61–87 (2014). https://doi.org/10.1007/s11721-014-0091-5
Set the name of the input image along with parameters of the calibration grid
in input.py. Now run correspondences_from_picks.py. If using ROS, you could
potentially use correspondences_from_tags.py from AprilTags detected in a
camera or stored image, but this file has not been touched in many years. The
result is known_correspondences.csv.
Next, run interpolator.py. It produces the file interpolated_correspondences.csv.
First written around April 2014. Updated in April 2025.
Andrew Vardy bots.cs.mun.ca