Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!asuvax!enuxha!nwatson From: nwatson@enuxha.eas.asu.edu (Nathan F. Watson) Newsgroups: comp.graphics Subject: Re: triangulation and contouring Message-ID: <1317@enuxha.eas.asu.edu> Date: 15 Aug 90 22:46:27 GMT References: <1990Aug15.003127.22609@NCoast.ORG> Organization: Arizona State Univ, Tempe Lines: 34 In article <1990Aug15.003127.22609@NCoast.ORG>, atul@NCoast.ORG (Atul Parulekar) writes: > I think there was a discussion some time back about delaunay triangulation of > random points. I wrote a program to do this based on a paper by Cavendish JC, > ... > Finally, any suggestions on how this triangle network can be used for > contouring directly (i.e., without first generating a grid of regularly > spaced points)? A student here at Arizona State University recently completed his master's thesis: "Contouring Trivariate Surfaces", Brett Keith Bloomquist. The thesis describes methods for contouring bivariate and trivariate surfaces (f(x,y) and f(x,y,z)). I believe a paper will eventually be published. For the bivariate case, the scheme boils down to: (a) Triangulate the data points, and estimate the gradients of the function at each data point. (b) Use the stuff obtained in (a) to obtain a Clough-Tocher interpolant over each triangle, resulting in a number of triangular Bezier patches of degree 3, which you will want to contour. (c) Approximate each degree 3 patch with one or more degree 2 patches (according to some user-defined tolerance). The degree 2 patches describe quadratic functions, which may be contoured using at most six rational quadric bezier curves per patch. The scheme is too detailed to explain fully here. I will attempt to let those interested know when the article appears. --------------------------------------------------------------------- Nathan F. Watson Arizona State University nwatson@enuxha.eas.asu.edu Computer Science Department