[vtkusers] trouble with vtkSurfaceReconstructionFilter

[Moore, Brett]

Hi,
I have a large dataset obtained from a laser scanner (approx 550K points).  I have experimented with several vtk rendering methods, but I have found nothing seems to work well.  The dataset represents a 360 degree scan of a statue (resolution is on the order of 1 cm).  Since it is a surface scan, vtkDelaunay3D does not work well.  vtkDelaunay2D fails because the dataset represents all sides of the object.
 
vtkSurfaceReconstructionFilter looks to be the best bet (especially since I have tried some of Hoppe's code on the data with encouraging results.  Unfortunately, Hoppe's 92 dissertation code is not maintained and does not scale well to the size of this dataset).  With the default sample spacing, vtkSurfaceReconstruction filter returns a recognizable rendering of the object - but it is sampled much too coarsely.  I expected the SetSampleSpacing method to correct this.  When I specify a sample size, my program simply dies with an "aborted" message after thinking for some time.  This behavior occurs regardless of the value I specify for sample size (my initial guess was 0.01, since all measurements are in meters).
 
Thus, my question is: can anyone shed some light on this behavior?  What causes the abort?  I am using RH 9, gcc 3.2.2, vtk 4.2.2.  Or...can anyone recommend a better approach?
 
Thanks,Brett