Tweet
I'm working on the calculation of horse-thoracic volume from collected
kinematic Data. This leaves me with a Point Cloud of 86 markers attached to
the body surface - which should now be triangulated and the volume
calculated.
The whole program is supposed to be written in Matlab.
So far I managed to do a triangulation as well as building a hull, using
Delaunay tessellation and the convhull function, both using the often used
"qhull" function (www.qhull.org).
Working fine on the lower parts of the Abdomen, since the surface curvature
is nicely convex, this approach leaves me with a big problem at the spine,
and the upper sides of the body - which are slightly concave! Qhull just
ignores some of the markers - to make the resulting surface as small as
possible.
Is there any way of assuring that all of the markers (points) HAVE to be
part of the resulting surface?
Or have you encountered a totally different, yet more satisfying approach?
Any suggestion is highly appreciated,
Thanks in advance,
Alexander Kotschwar
-----------------------------------------
Movement Science Group Vienna, Austria
http://msg.vu-wien.ac.at
kinematic Data. This leaves me with a Point Cloud of 86 markers attached to
the body surface - which should now be triangulated and the volume
calculated.
The whole program is supposed to be written in Matlab.
So far I managed to do a triangulation as well as building a hull, using
Delaunay tessellation and the convhull function, both using the often used
"qhull" function (www.qhull.org).
Working fine on the lower parts of the Abdomen, since the surface curvature
is nicely convex, this approach leaves me with a big problem at the spine,
and the upper sides of the body - which are slightly concave! Qhull just
ignores some of the markers - to make the resulting surface as small as
possible.
Is there any way of assuring that all of the markers (points) HAVE to be
part of the resulting surface?
Or have you encountered a totally different, yet more satisfying approach?
Any suggestion is highly appreciated,
Thanks in advance,
Alexander Kotschwar
-----------------------------------------
Movement Science Group Vienna, Austria
http://msg.vu-wien.ac.at