CGAL users discussion list

[Marco Feder <>]

Dear all,

I have two cubes (they’re coming from two cells of a 3D finite element mesh) and a square (a 2D object) between the two, intersecting both of them. I triangulated each cube with a Triangulation_3, and the square with a Delaunay_triangulation_2.

To compute the intersection of the square with the first cube, I do the following: I loop over all tets of the triangulated cube, and for each Tetrahedron I compute the intersection with a Triangle_3 coming from the triangulated 2D square. Oftentimes I have correct results, but in some scenarios I observe inaccurate results.

For instance: for each cube, I computed the area of the part of square inside it. The sum of these two areas should be precisely equal to the area of the square (it is exactly 0.136900), but it’s not. In particular, the first area is 0.0237738, and the second 0.0948123. Then the sum is not 0.136900.

The following GitHub gist shows this: https://gist.github.com/fdrmrc/49f397a5f72a7a00b1dfb992ab7babe1 . The output is the area of the intersection of the square with the second cube. To get the first area, change at line 52 pts_second_hex with pts_first_hex.

I can’t spot what is going wrong here.

Best,

Marco