On Wed, 7 May 1997, Darrell Goertzen wrote:
>
> PART 2. We will be using an Optotrack 3020 (active infrared markers)
> to record rigid body motion during testing in the above mentioned chamber.
> Therefore, a clear plastic? glass? barrier will be between the camera and
> the markers. We have not yet done a calibration check under these conditions
> and anticipate some kind of effect due to this barrier.
>
> The only similiar reference I have found is a paper by Lam et al.,
> (Calibration Characteristics of a Video Dimension
> Analyser (VDA) System, J Biomech 25(10):1227-1231 1992) which assessed the
> calibration of a 2D visible light measurement system with a 6.4mm thick,
> low-refractive index glass between the camera and specimen. They found
> changes in the calibration parameters with the glass and also by varying the
> angle of incidence.
>
> I would be happy to hear from anyone with experience using an Optotrack (or
> other active infrared system) with a clear sheet between the cameras and
> markers, or thoughts about distortion of 3D measurements, best type of
> material, etc.
Darrell:
Just a quick note to help you with your "barrier"
problem.
If you arrange things properly, your "barrier" can
become simply another element in the lens of your
camera, but a material with uniform optical
properties is essential if you intend to make any
useful correction. Cheap plastics do not qualify.
However, if you choose a reasonably good quality
glass, the effects of the refraction will be
consistent and uniform.
If the optical axis of your lens(es) are aligned
orthogonal to the glass, the effect produced by the
"additional element" can be modelled as pure radial
distortion. If, on the other hand, the optical
axis of the lens(es) strikes the glass surface
at an oblique angle, then the effect should be
modelled with both radial and tangential
components.
There is a relatively small body of literature dealing
with models of both radial and tangential distortion.
As a start, you might want to take a look at some
of the references cited in my doctoral dissertation:
Walton, James S. "Close Range Cine-Photogrammetry:
A Generalized Technique for Quantifying Gross Human
Motion". Doctoral Dissertation, The Pennsylvania
State University, 1981.
I'm traveling at the moment and don't have access
to additional details. If you drop me a note in
a couple of weeks, I'll see if I can find some
additional material for you.
As a postscript, for others on the biomechanics list,
anytime an air/water, air/glass interface is
introduced into the path of optical rays, the
effect can be modelled as optical distortion.
I've added this footnote, because I think the
approach might be helpful for those involved
with 3-D analyses of swimmers.
I've found that the best approach under these
circumstances is to first "map" the image-forming
properties of each "lens" (i.e., all refractive
surfaces involved in the view) and then go on to
the actual 3-D reconstruction.
Hope this helps.
Sincerely,
Jim Walton
Chair-Elect (1998-2000)
SPIE Working Group on High-Speed Photography,
Videography and Photonics.
************************************************** ****
************************************************** ****
* * *
* JAMES S. WALTON, Ph.D. * INTERNET: Jim@4DVideo.com *
* President * *
* 4D VIDEO * *
* 825 Gravenstein Hwy N. * PHONE: 707/829-8883 *
* Suite #4 * *
* SEBASTOPOL, CA 95472 * FAX : 707/829-3527 *
* * *
************************************************** ****
************************************************** ****
>
> PART 2. We will be using an Optotrack 3020 (active infrared markers)
> to record rigid body motion during testing in the above mentioned chamber.
> Therefore, a clear plastic? glass? barrier will be between the camera and
> the markers. We have not yet done a calibration check under these conditions
> and anticipate some kind of effect due to this barrier.
>
> The only similiar reference I have found is a paper by Lam et al.,
> (Calibration Characteristics of a Video Dimension
> Analyser (VDA) System, J Biomech 25(10):1227-1231 1992) which assessed the
> calibration of a 2D visible light measurement system with a 6.4mm thick,
> low-refractive index glass between the camera and specimen. They found
> changes in the calibration parameters with the glass and also by varying the
> angle of incidence.
>
> I would be happy to hear from anyone with experience using an Optotrack (or
> other active infrared system) with a clear sheet between the cameras and
> markers, or thoughts about distortion of 3D measurements, best type of
> material, etc.
Darrell:
Just a quick note to help you with your "barrier"
problem.
If you arrange things properly, your "barrier" can
become simply another element in the lens of your
camera, but a material with uniform optical
properties is essential if you intend to make any
useful correction. Cheap plastics do not qualify.
However, if you choose a reasonably good quality
glass, the effects of the refraction will be
consistent and uniform.
If the optical axis of your lens(es) are aligned
orthogonal to the glass, the effect produced by the
"additional element" can be modelled as pure radial
distortion. If, on the other hand, the optical
axis of the lens(es) strikes the glass surface
at an oblique angle, then the effect should be
modelled with both radial and tangential
components.
There is a relatively small body of literature dealing
with models of both radial and tangential distortion.
As a start, you might want to take a look at some
of the references cited in my doctoral dissertation:
Walton, James S. "Close Range Cine-Photogrammetry:
A Generalized Technique for Quantifying Gross Human
Motion". Doctoral Dissertation, The Pennsylvania
State University, 1981.
I'm traveling at the moment and don't have access
to additional details. If you drop me a note in
a couple of weeks, I'll see if I can find some
additional material for you.
As a postscript, for others on the biomechanics list,
anytime an air/water, air/glass interface is
introduced into the path of optical rays, the
effect can be modelled as optical distortion.
I've added this footnote, because I think the
approach might be helpful for those involved
with 3-D analyses of swimmers.
I've found that the best approach under these
circumstances is to first "map" the image-forming
properties of each "lens" (i.e., all refractive
surfaces involved in the view) and then go on to
the actual 3-D reconstruction.
Hope this helps.
Sincerely,
Jim Walton
Chair-Elect (1998-2000)
SPIE Working Group on High-Speed Photography,
Videography and Photonics.
************************************************** ****
************************************************** ****
* * *
* JAMES S. WALTON, Ph.D. * INTERNET: Jim@4DVideo.com *
* President * *
* 4D VIDEO * *
* 825 Gravenstein Hwy N. * PHONE: 707/829-8883 *
* Suite #4 * *
* SEBASTOPOL, CA 95472 * FAX : 707/829-3527 *
* * *
************************************************** ****
************************************************** ****