[vtkusers] Fwd: ICP Algorythm and parametric surfaces

[Elias Breunig]

---------- Forwarded message ----------
From: Elias Breunig <elias.breunig at googlemail.com>
Date: 2010/8/23
Subject: Re: [vtkusers] ICP Algorythm and parametric surfaces
To: David Doria <daviddoria+vtk at gmail.com>


Hi,

I want to provide the surface as equation (like for a paraboloid
e.g.). So sampling it with very high resulution can be a method to
allow a first "coarse" alignment, but in the end i need a fully
contious parametrized surface. I did a first version of my code in
matlab where i calculated the closest point set each iteration from
the parametric surface by normal shooting a initial guess and
consecutive gradient search starting from that point on
paraboloid/ellipsoid surface. This works quite fine but i was
wondering whether i would be able to use some of VTK analysis filters
to keep the whole project more compact and speed it up a little bit no
that I'm using VTK any way (vor the graphics output).

Any more ideas?

Thanks

Elias

2010/8/23 David Doria <daviddoria+vtk at gmail.com>:
> On Mon, Aug 23, 2010 at 5:01 AM, Elias Breunig
> <elias.breunig at googlemail.com> wrote:
>>
>> Hello,
>>
>> I'am relativeley new to vtk and want to use it for a small surface
>> analysis tool.
>>
>> Is there a straight forawrd way to link the ICP Filter to a parametric
>> surface as its registration target. As far as I understand the Docs,
>> only Point or Cell kind of data can be used, but it would surprise if
>> there is no workaround out there to use a surface.
>>
>> Thanks for your advice!
>>
>> Elias
>>
>
> What format is your surface currently in? I think you'll need to use
> vtkContourFilter (as in this example
> http://www.vtk.org/Wiki/VTK/Examples/Cxx/ImplicitFunctions/SampleFunction )
> to generate a sampled version of your surface before giving it to ICP.
>
> David