[vtkusers] what about speed? delaunay, decimate and QuadricClustering

[Will Schroeder]

If your mesh size gets very large you might consider using 
vtkQuadricClustering because you can process the data in pieces, i.e., you 
don't have to hold the entire dataset in memory. It's also quite fast.
Will


At 09:20 PM 3/19/2002 +0100, Alexander Trum wrote:
>Hello
>
>I'm having a polygonal model from car having 2,2 million triangles.
>I'm using a AMD Athlon 650 MHz with Linux, 500 MB RAM.
>To read in the model from STL ASCII format,
>to decimate the model with vtkDecimatePro
>and to write the decimated model in STL ASCII format,
>it takes round about 6 minutes.
>But the most time is required for reading and writing.
>With vtkDecimatePro i have got the best quality.
>
>Bye,
>   Alex
>
>"Sébastien MARAUX" schrieb:
> >
> > Hello,
> >
> > I want to triangulate a DEM with a file size of 4000 x 4700
> > (18 Million points with half of them having a value -> 9
> > millions points to handle)
> >
> > I am using a PIII 800 with 512 Mo SDRAM.
> >
> > I expect a speed of about 10 000 triangles / sec for this
> > part, is it OK ? (-> 15 minutes for this operation)
> >
> > I also want to reduce the number of triangles issued of
> > this DEM to about 10 000 or 100 000, using Decimate or
> > QuadricClustering functions.
> >
> > What time would it take to reduce 9 million triangles to
> > 10 or 100 thousands keeping a fairly good visual quality?
> >
> > I hope to achieve this operation in one night max, will it
> > be possible ?
> >
> > Sebastien MARAUX, total newbie in VTK functions exploration
> > and use.
> > ______________________________________________________
> > Boîte aux lettres - Caramail - http://www.caramail.com