home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.theory
- Path: sparky!uunet!munnari.oz.au!metro!usage!usage.csd!lambert
- From: lambert@spectrum.cs.unsw.oz.au (Tim Lambert)
- Subject: Re: comp geometry
- In-Reply-To: bross@splatter.nas.nasa.gov's message of Mon, 7 Dec 92 21:51:44 GMT
- Message-ID: <LAMBERT.92Dec15202435@nankeen.spectrum.cs.unsw.oz.au>
- Sender: news@usage.csd.unsw.OZ.AU
- Nntp-Posting-Host: nankeen.spectrum.cs.unsw.oz.au
- Organization: CS&E, Uni of NSW, Australia
- References: <1992Dec7.215144.962@nas.nasa.gov>
- Date: Tue, 15 Dec 1992 10:24:35 GMT
- Lines: 25
-
- >>>>> On Mon, 7 Dec 92 21:51:44 GMT, bross@splatter.nas.nasa.gov (Bill Ross) said:
-
- > Does anyone have a formula or algorithm for calculating the biggest
- > sphere that can fit in an irregular hexahedron?
-
- If the hexahedron is convex, then the sphere will touch four faces, so
- you could test all 6_choose_4 subsets of the faces to see if the
- sphere touching the four support planes is inside the hexahedron and
- take the largest such sphere.
-
- If it is not convex then you have to worry about spheres tangent to
- concave vertices and edges.
-
- > How about an n-sided figure?
-
- If it is convex, you could take the dual, and find the smallest
- enclosing sphere by the obvious generalization of of the smalles
- enclosing circle algorithm.
-
- If not convex, then the centre of the sphere lies on Voronoi diagram
- of the polyhedron, so searching that should do it, but I don't know of
- a published algorithm for this...
-
- Tim
-
-