modified question

In the following example

pr = RandomReal[{0, 1}, {100, 2}];
p = Select[pr , #[[1]] #[[2]] < .2 &];
Show[{ DelaunayMesh[p], ListPlot[p, PlotStyle -> Black], Plot[.2/x, {x, 0, 1}, PlotRange -> {{0, 1}, {0, 1}}]}]

enter image description here

meshing (same with ConvexHullMesh) shows several elements above the hyperbolean which I would like to remove.

How to solve this problem for an arbitrary point set with Mathematica?


  • 1
    $\begingroup$ But one can still cut out some triangles from your last plot to reduce the area even more. And maybe make it = 0. What is the exact mathematical definition of the hull you are looking for? Notice that fig. 2 does not represents a convex hull. $\endgroup$ – yarchik Feb 21 at 15:55
  • $\begingroup$ You're right, the example is to simple. In my underlying problem I need to discretize a region (something like a triangle with a concave side) for further interpolation. Convexhull of this region expands the region unacceptable. Thanks for your interest! $\endgroup$ – Ulrich Neumann Feb 21 at 17:34
  • $\begingroup$ @yarchik I'll modify my question... $\endgroup$ – Ulrich Neumann Feb 21 at 17:44
  • $\begingroup$ Relevant keywords are "concave hull" and "α-concave hull". $\endgroup$ – yarchik Feb 22 at 18:35
  • 1
    $\begingroup$ @yarchik Appears to be promising, I'll test it. $\endgroup$ – Ulrich Neumann Feb 23 at 10:46

Browse other questions tagged or ask your own question.