Tweet
This interesting discussion on DLT calibration assumes (I think) that
the same calibration dataset is used for estimating camera/lens
nonlinearity (i.e. distortion) as well as the basic 11 DLT parameters.
It makes sense that many control points are needed in this case. This
method also requires that the mathematical model chosen for distortion
modeling is a good fit for the actual distortion pattern. Deviation of
actual distortion from the model is likely near the edges of the field,
which can create the extrapolation errors we are all familiar with.
Some methods employ a seperate procedure using a uniform marker grid
(typically consisting of several hundred markers covering the entire FOV
of the camera) to generate a distortion correction mapping specific to a
given camera/lens combination. This mapping info is then used to remove
nonlinearities from the 2D images (usually using trilinear interpolation
from the grid data) before DLT calibration. It would seem that the
separate distortion correction procedure, if implemented properly, would
reduce the number of control points required for good accuracy and
improve calibration near the edges of the field of view.
I have never seen a comparison of the two different methods. Does
anyone have any thoughts on the relative merits of the one-step vs
two-step processes?
__________________________________________________ ___________________
Scott Tashman, Ph.D.
Head, Motion Analysis Section Assistant Professor
Bone and Joint Center Department of Orthopaedics
Henry Ford Hospital School of Medicine
2799 W. Grand Blvd. Case Western Reserve University
Detroit, MI 48202
Voice: (313) 876-8680 or 876-7572
FAX: (313) 556-8812 or 876-8064
Internet: tashman@bjc.hfh.edu
__________________________________________________ ___________________
the same calibration dataset is used for estimating camera/lens
nonlinearity (i.e. distortion) as well as the basic 11 DLT parameters.
It makes sense that many control points are needed in this case. This
method also requires that the mathematical model chosen for distortion
modeling is a good fit for the actual distortion pattern. Deviation of
actual distortion from the model is likely near the edges of the field,
which can create the extrapolation errors we are all familiar with.
Some methods employ a seperate procedure using a uniform marker grid
(typically consisting of several hundred markers covering the entire FOV
of the camera) to generate a distortion correction mapping specific to a
given camera/lens combination. This mapping info is then used to remove
nonlinearities from the 2D images (usually using trilinear interpolation
from the grid data) before DLT calibration. It would seem that the
separate distortion correction procedure, if implemented properly, would
reduce the number of control points required for good accuracy and
improve calibration near the edges of the field of view.
I have never seen a comparison of the two different methods. Does
anyone have any thoughts on the relative merits of the one-step vs
two-step processes?
__________________________________________________ ___________________
Scott Tashman, Ph.D.
Head, Motion Analysis Section Assistant Professor
Bone and Joint Center Department of Orthopaedics
Henry Ford Hospital School of Medicine
2799 W. Grand Blvd. Case Western Reserve University
Detroit, MI 48202
Voice: (313) 876-8680 or 876-7572
FAX: (313) 556-8812 or 876-8064
Internet: tashman@bjc.hfh.edu
__________________________________________________ ___________________