CGAL users discussion list

["Santosh Tiwari" <>]

Thank you for the informative email.

I am using 3D alpha shapes, and it does what it is supposed to - a
non-convex hull of the point cloud. It however cannot be used to create a
smooth surface from the point cloud. It exactly interpolates the point
cloud data (that is what it is supposed to do). It also generates many
artifacts inside the outer surface as if those were part of the model. I
am currently writing my own code to generate a mesh from very noisy point
cloud data; more than 70% of the points are inside the volume of the
object, i.e. they should not be used while generating the surface mesh.



> Ok, now I found a way to remove the facets that I didn't want - those
> facets closing the surrounded space.
> I just inserted a call to "is_Gabriel" for selection.
>
> I also found another way I could have tried if that would not have
> worked - 3D-Alpha-Shapes, but because it worked I didn't try...
>
> Maybe when my collegue tried some other point-sets and has any problems
> with this current solution, or if I should ever have plenty of time and
> nothing to do with that...
>
> With best regards,
>
> Sven
>
> Sven Andres schrieb:
>> Hello everybody,
>>
>> since my first message ("[cgal-discuss] non-convex Objects from
>> Pointset") I did some more reading and trying and I found, that it
>> does not seem to change the shape of the resulting object if I just
>> change the triangulation-method. Maybe the created faces are slightly
>> different, but still the result is a closed model and not the surface
>> I'm looking for.
>>
>> For those who did not read my first request - I try to turn a set of
>> points (with free orientation in space) into a 3D-Model of the surface
>> described by the points - e.g. just a patch of a spheres outer hull,
>> not a closed object as though it was cut out of a massive sphere. I
>> got to a point where I can extract a set of faces describing the
>> closed object but that is not what I want... A possible way would be
>> to try to project the point-set onto a plane and do 2D-triangulation
>> but we did it like that in the past and we wanted to go away from
>> 2D-triangulation for we got errors on the boundaries with that approach.
>>
>> My collegue told me he knows about a program (geomagic) using CGAL
>> that offers an "open 3D triangulation" so that you can turn a set of
>> points describing a surface into a non-closed model of the described
>> surface. He suggested that this might work when using another
>> triangulation method or a different geometry-kernel, but I could not
>> find a way to do this yet...
>>
>> Since yesterday I went away from "triangulation" and read something
>> about ways to handle 3D-surfaces rather than bodies in CGAL.
>>
>> I read about "envelopes" and "surface_meshing". While surface-meshing
>> seemed like a step into the right direction it looks like it cannot be
>> used on a free set of points - without a complete description of the
>> surface, or at least one giving the possibility to interpolate between
>> given values.
>>
>> So here I go again. Can anybody give me a hint how to turn any set of
>> points measured from a surface into a set of faces describing the
>> surface?
>> Is it possible with CGALs "surface-meshing" or do I need to go a
>> different way?
>>
>> Any help is absolutely appreciated.
>>
>> With best regards,
>>
>> Sven Andres
>> No virus found in this incoming message.
>> Checked by AVG.
>> Version: 7.5.524 / Virus Database: 269.23.15/1426 - Release Date:
>> 10.05.2008 11:12
>>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss
>


-- 
Santosh Tiwari
Fluor Daniel EIB 326
Clemson University, Clemson, SC 29634
http://www.clemson.edu/~stiwari/