[vtkusers] Probing unstructured point data onto an unstructured grid

[Prabhu Ramachandran]

>>>>> "GL" == Goodwin Lawlor <goodwin.lawlor at ucd.ie> writes:

    GL> Hi Alex, A couple of things might work:

    GL> 1. You could use either vtkShepardMethod or
    GL> vtkGaussianSplatter to sample the brick_8.ipr data onto a
    GL> volume. You could then probe this volume with the brick_8.vtk
    GL> data

    GL> 2. It looks like the brick_8.ipr could be topologically
    GL> regular. If thats the case, then use vtkStructuredGrid to hold
    GL> your brick_8.ipr data. You could then probe the structured
    GL> grid.

One more point to note.

The basic problem in your (Alex's) approach is that you are trying to
probe a set of points (which have scalars corresponding to your
solution) with a volume.  Whichever way you probe it it you will only
get points and not volumes.  The way around it is to make your
block_8.ipr into a volume.

You (Alex) also mention that the brick_8.ipr is the result of a finite
element integration.  Clearly this finite element integration will be
performed inside some set of volumes (an unstructured grid in itself)
and you will have the solution either at the nodes or the cells of
these volumes.  Why not use these volumes to do the job directly
rather than just obtaining the data at various points and then sending
the points alone through ShepardMethod/GaussianSplatter and then
obtaining another approximated volume?

So if you have the unstructured grid from your FEM simulation use that
and probe it with your brick_8.vtk unstructured grid.  

If this is not an option and you dont have the unstructured grid from
the FEM simulation you need to take the approach specified by Goodwin.

cheers,
prabhu