Radial distortions can be modeled through
polynomial, division and rational functions. Points based
solutions to recover the coordinate of the principal point
and the coefficients of radial lens distortions though
polynomial, division and rational models, have been
proposed in the literature. Points based calibration
strategies have been criticized for the lack of accuracy of
the calibrated camera parameters. Line-based
calibration approaches have been hailed for their good
accuracy due to strong geometric constraints imposed on
calibration points such as the collinearity constraints.
However, in the presence of severe radial distortions due
to lens imperfections some line-based calibration
strategies fail to model with accuracy the curvature of
the 3D world resulting from severe lens distortions. This
paper presented a new radial distortion calibration
theory based on some properties of hyperbolic curves.
The developed and proposed mathematical algorithms
enable to recover the parameters of radial distortion as
well as the coordinates of the principal point without any
need for iterations or minimization of a cost function.
Keywords : Camera Calibration, Principal Point, Radial Lens Distortions.