CGAL users discussion list

[jpanetta <>]

Dear CGAL developers,

I'm trying to mesh a domain defined by a signed distance function with sharp
features. The signed distance function represents the intersection of a
smooth surface with the unit cube [0, 1]^3. My approach is to extract the 1D
sharp features using marching squares on the 6 cube faces, then call
make_mesh_3 as in the "Mesh_3/mesh_two_implicit_spheres_with_balls.cpp"
example.

I am able to get good results by playing with the marching squares grid size
and CGAL's meshing criteria, but in general I experience two problems:

1) Many marching squares grid sizes produce feature curves that trigger the
following assertion:

CGAL error: assertion violation!
Expression : n==0 || dleft_frac >= float_prior(float_prior(1.))
File       : /opt/local/include/CGAL/Mesh_3/Protect_edges_sizing_field.h
Line       : 1025
Explanation:
Refer to the bug-reporting instructions at
http://www.cgal.org/bug_report.html
libc++abi.dylib: terminating with uncaught exception of type
CGAL::Assertion_exception: CGAL ERROR: assertion violation!
Expr: n==0 || dleft_frac >= float_prior(float_prior(1.))
File: /opt/local/include/CGAL/Mesh_3/Protect_edges_sizing_field.h
Line: 1025

The feature curves all look fine to me; some vertices are close together,
but there are no degenerate edges. Is there a minimal edge length that I
must satisfy?

2) For the marching squares grids that work, setting the edge_size criterion
"too low" causes the program to hang after printing (with
CGAL_MESH_3_VERBOSE defined):

Scanning triangulation for bad facets (sequential) - number of finite facets
= 3034...

The finer marching squares grids let me set this criterion lower, but I
haven't yet been able to determine the precise relationship between the
feature curve edge lengths and the lower bound on edge_size that triggers
this hang. Is this expected behavior?

Thanks,
Julian



--
View this message in context: 
http://cgal-discuss.949826.n4.nabble.com/Some-3D-Mesh-Generation-Problems-Sharp-Features-tp4660976.html
Sent from the cgal-discuss mailing list archive at Nabble.com.