> Date: Wed, 1 Feb 2012 14:09:47 -0500 > From: Raul Miller <rauldmil...@gmail.com> > Subject: Re: [Jprogramming] planar Delaunay and Voronoi available > To: Programming forum <programming@jsoftware.com> > Message-ID: > <cad2jou_dbnw52uld0mwwvjbwth2w5t_hbvu5z83rnti4dkp...@mail.gmail.com> > Content-Type: text/plain; charset=UTF-8 > > On Wed, Feb 1, 2012 at 12:07 PM, David Ward Lambert > <b49p23t...@stny.rr.com> wrote: > > http://www.jsoftware.com/jwiki/DelaunayAlgorithms > > Do you have any available sample data sets for exercising this code? > > Thanks, > > -- > Raul
0) Yes, below. 1) I don't have a proof that my Voronoi algorithm is correct. 2) This data set discloses an implementation bug in the voronoi code which I'll attempt to fix soon. 3) The delaunay mesh looks correct. As compared to the picture it is flipped vertically and hasn't got the edges of the convex hull around the data set. I'll investigate a more careful discard of the bounding box. I used gimp to flip the picture. Multiplying the data looks easier. 1 _1 *"1 p Test Figure: www.cs.mtu.edu/~shene/PUBLICATIONS/2004/Hull2VD.pdf page 9 Guesstimated data with left handed display coordinate system: p=: ".;._2]0 :0 454 392 251 702 621 668 633 453 423 616 224 518 419 803 552 535 356 436 400 184 688 233 724 491 724 777 128 755 78 453 153 211 85 57 497 129 765 70 793 458 779 824 65 810 146 592 17 192 197 392 ) plot_tria_mesh mesh=: triangulate p voronoi p NB. seems to work with y as 6 2 ?.@$ 0 |length error: convex_hull | y=.y ([/:angle@:-"1)start ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
