[vtkusers] eliminate duplicate nodes

David E DeMarle dave.demarle at kitware.com
Thu Dec 6 12:53:30 EST 2007

On 12/6/07, burlen <burlen at apollo.sr.unh.edu> wrote:
>  I see where you are coming from, vtk is awesome in that it provides a lot
> of well documented very thoroughly debugged functionality. Just a thought...
> vtkPolyData is a type(not in the C++ sense) of unstructured grid, in fact
> all of the cell types applicable for use with the unstructured grid are
> available for use with the polydata. In my experience vtkPolyData class
> encapsulates the functionality of the vtkUnstructuredGrid, although they are
> not symetrically interchangable. You may be able to use poydata in place of

You have this point reversed. PolyData cell types are a subset of
those in UnstructuredGrid. PolyData, is tuned for graphics cards and
rendering. PolyData can only represent 0 (point), 1 (line segment) and
2 (polygons, pixels etc) dimensional cells. UnstructuredGrid can also
represent 3 dimensional cells (tetrahedra, voxels etc).

> the unstructured grid. & If you couldn't it is very easy to convert between
> the two, by taking the point set, cells, and point or cell data from one and
> inserting it into the other. So what you might try is to convert your
> unstructured grid to polydata clean it then convert back.
GeometryFilter will extract the surfaces of all boundary cells in an
UnstructuredGrid, producing PolyData. I am not aware of any filters
that can reverse that process (taking the 4 triangles that are on the
surface of a tetrahedra and producing the tetrahedra for example.)

>  Burlen
>  Oliver Gloth wrote:
>  I had a look at vtkCleanPolyData and the classes vtkMergePoints and
> vtkPointLocator seem to be good starting points. It is required to clean a
> vtkUnstructuredGrid, that's why the vtkCleanPolyData is not directly
> applicable - but thanks for pointing me in the right direction.
> The links you sent (thanks btw.) are not completely unknown to me and I
> might use the Netgen package for this project. It is not a CFD project, but
> a mesh generation project (I would not use VTK data structures for a solver
> ...). I gave some thought to using VTK: it might not be ideal, but
> definitely well suited. Tools like, for example, vtkMergePoints, are very
> valuable and they can be a real pain to implement and debug. The other thing
> I like about VTK is that it is well documented and has a large user
> community, and yes visualisation is very easy indeed!
> Oliver
> -----Original Message-----
> From: burlen [mailto:burlen at apollo.sr.unh.edu]
> Sent: 06 December 2007 16:28
> To: Oliver Gloth
> Cc: vtkusers
> Subject: Re: [vtkusers] eliminate duplicate nodes
> Does the vtkCleanPolyData class do what you want??
> Although I supose you could do it, I am not convinced that VTK would be
> ideal for CFD project, you might find these links of use:
> http://www.andrew.cmu.edu/user/sowen/softsurv.html
> http://www.cs.cmu.edu/~quake/triangle.html
> http://www.hpfem.jku.at/netgen/
> On the other hand at least visualization would be easy.
> Burlen
> Oliver Gloth wrote:
>  Dear VTK Users,
> I am currently working on an open-source project - developing a mesh
> generator for CFD simulations. Due to budget and time constraints I have
> chosen to use the VTK data structures as a basis, instead of writing my
>  own
>  data structures. Furthermore this makes building a GUI (based on Qt) to
> modify and visualise the mesh easy. One thing that is required is a filter
> to merge two grids and eliminate duplicate nodes. Is such a thing already
> part of VTK, or does somebody have a piece of code that would do this? The
> implementation is not trivial, because the 3D search can be very time
> consuming. If it doesn't exist yet I will implement it myself, but I would
> like to avoid double work as much as possible.
> Oliver
> _______________________________________________
> This is the private VTK discussion list.
> Please keep messages on-topic. Check the FAQ at:
> http://www.vtk.org/Wiki/VTK_FAQ
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers

More information about the vtkusers mailing list