Re: irregular data

• To: John Yao <johnyao@xxxxxxxx>
• Subject: Re: irregular data
• From: Bill Hibbard <billh@xxxxxxxxxxxxx>
• Date: Fri, 12 Jul 2002 16:17:44 -0500 (CDT)

Hi John,

> I am currently dealing with the problem of generating 3D shapes from an
> "unknown" (in terms of values) set of data.  Please allow me to explain:
>
> 1.  I would like to sample the all of the points in a sphere that is
> generated by a method shown in a pic at the very bottom of the following
> webpage: http://astronomy.swin.edu.au/~pbourke/modelling/sphere.
>
> 2.  Each of the sampled points will be used to compute a value.
> 3.  These values will be used to form a new surface
> (shape).  Unfortunately, we cannot anticipate the geometry of this shape.
>
> How should I go about generating this initially unknown geometry?  I was
> thinking of feeding the computed values into the Irregular3DSet and somehow
> use the Delaunay functions, but really I don't have any idea of how these
> two things work?  Can someone give me some pointers?

Since you say surface, you need an Irregular3DSet with manifold
dimension = 2. Use the Irregular3DSet constructor with a Delaunay
argment, and pass it a DelaunayCustom. Construct that by passing
it the same 'float[][] samples' array of locations that you pass
to the Irregular3DSet constructor, plus an 'int[][] tri' array
of triangles (dimensioned [number_of_triangles][3] and containing
indexes of points in the samples array). You'll have to somehow
compute what those triangles should be, which may not be so hard
based on the images on your web page. Split rectangles into pairs
of triangles.

Good luck,
Bill

• References:
  • irregular data
    • From: John Yao