Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!mrsvr.UUCP!haletky
From: haletky@mrsvr.UUCP (Ed Haletky)
Newsgroups: comp.graphics
Subject: Re: contour plotting
Message-ID: <771@mrsvr.UUCP>
Date: 22 May 89 12:41:40 GMT
References: <4487@alvin.mcnc.org>
Distribution: na
Organization: GE Medical, MR Center, Milwaukee
Lines: 38

From article <4487@alvin.mcnc.org>, by pickel@mcnc.org (Lisa C. Pickel):
> Can anyone recommend some algorithms/books on producing contour
> plots?  I think I understand the "triangles method", but unfortunately
> my data is not on a regular, rectangular grid.
> 
> Code on which I could hack would also be more than welcome.
> 
> Thanks in advance,
> Lisa Pickel
> pickel@mcnc.org
> decvax!mcnc!pickel

You do not need a regular, rectangular grid to use the triangle method. The
data must form a triangle. Most of the time I have used the triangle method
I have been using 3d coordinates. I only concern myself with 2d of the 3d.
The third dimension can be ignored. Unless it is important to you. Here is the
example:

	I had a 3d non-rectangular, skewed mesh with a wave. That is how it looked
	physically. Computationally it was a cube. It is easy to map 3d points to
	a cube. Contour by triangles was then performed. I got some real nice
	contour plots.

I know this is not a very good explanation. But without looking through tons
of code it is all I have. I do have a half way decent 2d contour program,
written by myself and a professor from Purdue University.

Alternative methods would include: bezier curves
								   B-splines
								   B-spaline patches
								   Any other curve drawing technique.

					Good-luck,
Edward L. Haletky=================== =======Usenet:=========================
GE Medical Systems           |E   ^| |^   U|~{uwvax}!uwmcsd4!mrsvr!haletky 
3200 N. Grandview Blvd, W-826|L  / | | \  S|~{texsun|sunbird|sunbrew|
Waukesha, WI 53188           |H /__|||__\ A|        crdgw1}!gemed!haletky
==================================^===^=====================================