[vtkusers] radii of vtkCylinder and vtkCylinderSource are unequal?

[Maarten Beek]

Hi David, others,


I used the vtkButterflySubdivisionFilter class, which generated a much better result. However it also introduced an error (?), but it proved it was an accuracy issue and not my mistake.


Therefore, I created my own source which is working well.
I now want to implement an optional transformation to put the model anywhere I want.

So far I coded it like:

vtkPolyData* output = vtkPolyData::SafeDownCast( outInfo->Get(vtkDataObject::DATA_OBJECT));


< snip >

output->SetPoints( pts );
output->SetPolys( tris );


if ( this->Transform )
{
    vtkSmartPointer<vtkTransformPolyDataFilter> transformer = vtkTransformPolyDataFilter::New();
    transformer->SetInput( output );
    transformer->SetTransform( this->Transform );
    transformer->Update();

    transformer->SetInput( 0 ); // to disconnect input ?
    output = transformer->GetOutput();

}

In other words I try to pass the data from input to output to avoid temporarily copies of the data. (ITK has the DisconnectPipeline function for this). The code above however is not working.

The following code does work, however it uses this temporarily copy I try to avoid:

if ( this->Transform )
{
    vtkSmartPointer<vtkPolyData> temp = vtkPolyData::New();
    temp->ShallowCopy( output );

    vtkSmartPointer<vtkTransformPolyDataFilter> transformer = vtkTransformPolyDataFilter::New();
    transformer->SetInput( temp);
    transformer->SetTransform( this->Transform );
    transformer->Update();
    output->ShallowCopy( transformer->GetOutput() ); // Why is this shallow copy required?
 
}

Any suggestions?

Maarten



________________________________
 From: David Gobbi <david.gobbi at gmail.com>
To: Maarten Beek <beekmaarten at yahoo.com> 
Cc: "vtkusers at vtk.org" <vtkusers at vtk.org> 
Sent: Wednesday, December 7, 2011 5:03:12 PM
Subject: Re: [vtkusers] radii of vtkCylinder and vtkCylinderSource are unequal?
 

Hi Maarten,

VTK has lots of subdivision filters in the "Graphics" directory.  I would not advise
trying to change the way VTK does clipping.

Writing your own source is a good idea.  After you have done it once, it is easy
to do over and over again for whatever shape you need.  I have my own set of
special-purpose polydata sources and haven't used the ones that come with
VTK in ages.

 - David



On Wed, Dec 7, 2011 at 2:33 PM, Maarten Beek <beekmaarten at yahoo.com> wrote:

Thanks David,
>
>
>Good to know it is an approximating issue, and not me.
>
>
>However, I don't really know where the plane is. Although I could calculate it, but then I could position the vertices and triangulate between them myself (no need for the vtkClipPolyData class...).
>
>
>I guess improving the clipping would involve vtkGenericCell::Clip().
>
>
>Also, I made a simple application in which I can change the length of the cone relative to the radius of the cylinder; I don't see the clipping improve when I do this... so I am not yet convinced regarding your rule.
>
>
>Does VTK have a class that triangulates the triangles in a vtkPolyData? The cone resolution just makes the triangles narrower or wider (they keep going from bottom to top).
>
>
>I think that creating my own ConeCylinderSource class is easiest...
>
>
>Maarten
>
>
>
>________________________________
> From: David Gobbi <david.gobbi at gmail.com>
>To: Maarten Beek <beekmaarten at yahoo.com> 
>Cc: "vtkusers at vtk.org" <vtkusers at vtk.org> 
>Sent: Wednesday, December 7, 2011 3:26:42 PM
>Subject: Re: [vtkusers] radii of vtkCylinder and vtkCylinderSource are unequal?
>
>
>Refining the cone resolution will help, but if you can, you should use
>a plane to clip the cone, instead of using a cylinder.
>
>VTK does clipping by evaluating the implicit function at each end of
>each line segment, and then it does a linear interpolation along the
>line segment to find the point where it does the clip.  This works
>perfectly when the function is a plane function, but it is only
>approximate if the function is a curved surface.
>
>The rule is: the lengths of the polygon edges must be much less then
>the radius of curvature of the implicit function.  A plane has an
>infinite radius of curvature, so clipping with a plane is always
>ideal.
>
>- David
>
>
>On Wed, Dec 7, 2011 at 12:46 PM, Maarten Beek <beekmaarten at yahoo.com> wrote:
>> Hi all,
>>
>> See code below.
>> What I am trying to do is to connect a piece of cone to a cylinder. To do
>> this I cut the top of the cone using an implicit cylinder with the same
>> radius as real cylinder.
>> However, the radius of the cut cone top is not equal to the radius of the
>> real cylinder.
>> What am I doing wrong?
>>
>> Does this have to do with how the clip values in the points are
>> interpolated? And should I refine the triangulation of the cone to get a
>> better result?
>>
>> Thanks - Maarten
>>
>> vtkSmartPointer<vtkConeSource> startcone = vtkConeSource::New();
>>   startcone->SetResolution( 30 );
>>   startcone->SetHeight( this->EntryLength );
>>   startcone->SetRadius( this->EntryRadius );
>>  
 startcone->SetCenter( 0.0, -0.5*this->Length + 0.5*this->EntryLength, 0.0
>> );
>>   startcone->SetDirection( 0.0, 1.0, 0.0 );
>>   startcone->CappingOff();
>>
>>   vtkSmartPointer<vtkCylinderSource> cylinder = vtkCylinderSource::New();
>>   cylinder->SetResolution( 30 );
>>   cylinder->SetRadius( this->CylinderRadius );
>>   cylinder->SetHeight( this->Length );
>>   cylinder->SetCenter( 0.0, 0.0, 0.0 );
>>   cylinder->CappingOff();
>>
>>   vtkSmartPointer<vtkCylinder> cyl = vtkCylinder::New();
>>   cyl->SetRadius( this->CylinderRadius );
>>   cyl->SetCenter( 0.0, 0.5*this->Length, 0.0 );
>>
>>   vtkSmartPointer<vtkClipPolyData> clipper1 = vtkClipPolyData::New();
>>   clipper1->SetClipFunction( cyl );
>>  
 clipper1->SetInputConnection( startcone->GetOutputPort() );
>>
>>   vtkSmartPointer<vtkAppendPolyData> append = vtkAppendPolyData::New();
>>   append->AddInput( clipper1->GetOutput() );
>>   append->AddInput( cylinder->GetOutput() );
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20111215/2fc3889b/attachment.htm>