[SciPy-user] 3D (I'm not talking about rendering)

[Bryan Cole]

On Thu, 2008-01-17 at 10:54 +0100, José María García Pérez wrote:
> I have started to use TVTK, but just for rendering. How could I ask
> TVTK to tell me groups of nodes which are less than 0.005 far away
> ones from each others?
> 
> Could you write the name of a function that can do that task (or a
> similar task)? 

VTK (or TVTK) is indeed your friend. If you have 300000 points and you
want to find all occurences where the point spacing is below a certain
threshold you could use the "Gaussian Splatter" filter to find
approximate regions of high point density (see
http://www.vtk.org/doc/release/5.0/html/a01413.html or maybe the related
Sheppard Method). This would give you a measure of the point density on
a regular grid. You could sample this using your original points and
then filter them based on this point-density function. This would
hopefully reduce your point set to something small enough to do an exact
find-the-nearest-neighbour routine on (needs a N*N matrix, where N is
your number of filterd points). You could use a vtkPointLocator for this
but it's probably easier/faster to just write one with python/numpy.

If you're in no hurry, you could just apply the vtkPointLocator to all
300000 points.

> 
> Regards,
> Jose M.
> 
> 2008/1/17, jelle <jelleferinga at gmail.com>:
>         Have a look at TVTK (traited VTK ).
>         
>         The underlying Visualization ToolKit  is what drives some of
>         the largest
>         visualizations jobs around, and will surely grok that amount
>         of data.
>         Note that there's a very good guide on VTK, which I highly
>         recommend. 
>         All the operations you mentioned are well supported in VTK.
>         
>         -jelle
>         
>         _______________________________________________
>         SciPy-user mailing list
>         SciPy-user at scipy.org
>         http://projects.scipy.org/mailman/listinfo/scipy-user
> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user