Hi Roman,
The algorithm(s) I am most familiar with have to do with finding surfaces
of constant value (already) given a multi-resolution volume representation
(with the minimal number of cubes). The algorithm of volume tessellation
with the minimal number of cubes sounds a little bit different. Are you
interested in a multi-resolution volume data set representation in order
to minimize the number of cubes necessary to represent the volume? Or are
you interested in representing the outside surface of the volume with the
minimal number of cubes?
I'm not familiar with the HLA or DDM but it would not surprise me if
someone on the VisAD list might have some input into your inquiry.
-cheers, bob
Robert S Laramee tel: (603) 868-1361
9 Woodman Road, #313 office: (603) 862-0350
Durham, NH 03824 URL: http://www.cs.unh.edu/~rlaramee
On Mon, 14 Aug 2000, Roman Grigoriev wrote:
> Dear Sir!
> Could you be so kind to give me advice.
> I have a question on object representation by parallelepiped primitives
> My ph.d. work is in some way connected with volume approximation methods.
> I need an algorithm of volume tesselation with minimal number of cubes.
> The task is to solve intersection between sensor's field of view and
> target's signature in 3D.
> I think that the solution is in voxel representation of signatures and
> fields of view.
> This problem is actual in distributed simulation. The best solution for
> distributed simulation is HLA
> (High Level Architecture). But HLA is archtecture not toolkit, that's
> why developers should programm more and more.
> HLA has DDM- data distribution managment -service that route data
> between simulators.
> And this service is not sophisticated. so it supports only
> parallelepiped representation of fields of view. When parallelepiped of
> sensor FOV intersects with parallelepiped of target signature data will
> be transered between simulators. But in real life sensor field of view
> is cone-shaped and target signature is complex 3D shape. That's why the
> task is volume approximation with minimal number of cubes or
> parallelepipeds.
> So I ask for advice.
> Thanx ahead
> Bye
> Roman
