[vtkusers] Contouring through constant value areas of dataset

Burlen Loring burlen.loring at kitware.com
Sat Sep 6 14:10:43 EDT 2008


Ps:

I messed around a little with a test multiblock data set, basically 
contours terminate at block boundaries because to the vtk contouring 
algorithm blocks are not connected to each other, (ie cells adjacent 
across block boundaries do not share nodes, and thus are not connected 
to each other).

It looks like your input data the constant regions may be in their own 
blocks, which explains why you have contour lines terminating on their 
boundaries.

If all nodes of a dataset (or in your case block) have a constant value 
then it presents a degenerate case for the contouring algorithm because 
this configuration never results in a contour line intersecting any of 
the cell's edge's. In that case nothing would be drawn. See p 166 of the 
vtk text for a better explanation.

Burlen





Gerrick,

There is something suspect with the data you are contouring.
http://www.nabble.com/file/p19338704/constantValues-surface.png

Let me explain why I think this. Contouring uses interpolation along
cell edges, making use of nodal data. In my experience, when rendering
nodal data vtk interpolates so that in the rendered image you never see
a discontinuity across adjacent cells when adjacent cells share nodes
(it will always show as a steep but smooth transition). In your images
we can see discontinuities between the regions, therefor I wonder if
those cells are actually sharing nodes with their neighbors. ? In other
words to vtk they might not be connected regions.

Just a guess.
Of course it would be easier to help if you posted your data.

Burlen

David E DeMarle wrote:
> Ah, so presumably one of those isocontours happens to hit the
> contoured value exactly and you expect the whole thing to be drawn but
> it is not.
>
> You can reproduce the behavior with any Source, using Calculator to
> give every point the same constant scalar value, and then contouring
> on that value.
>
> You would expect everything to be drawn, but nothing is. I don't know
> why it does that.
>
> anyone?
>
> On Fri, Sep 5, 2008 at 4:38 PM, Gerrick Bivins
> <gbivins at objectreservoir.com> wrote:
>   
>> Right. But for instance, if I were visualizing this data via volume
>> rendering, the areas with this constant value would be displayed rather than
>> not.
>> Am I misunderstanding the intent of contouring? If so, is there another
>> filter I can use to accomplish what I'm describing?
>> Gerrick
>>
>>
>> On 9/5/08 3:29 PM, "David E DeMarle" <dave.demarle at kitware.com> wrote:
>>
>>     
>>> Umm, because the values are constant? Which means they contain only
>>> one value, and thus either nowhere contain the value that a contour
>>> displays, or everywhere contain it.
>>>
>>> I am not sure I understand your question otherwise.
>>>
>>>
>>> On Fri, Sep 5, 2008 at 4:17 PM, Gerrick Bivins
>>> <gbivins at objectreservoir.com> wrote:
>>>       
>>>> Hi All,
>>>> When generating contour values (isosurfaces) with vtkContourFilter, I've
>>>> noticed that areas in the dataset with constant values ( probably not the
>>>> correct terminology here) don't produce contours.
>>>>
>>>> For example, here is the original multiblock dataset in Paraview with areas
>>>> of constant value:
>>>> http://www.nabble.com/file/p19338704/constantValues-surface.png
>>>>
>>>> Here is the contouring with about 100 values:
>>>> http://www.nabble.com/file/p19338704/constantValues-100-contours.png
>>>>
>>>> Here is the contouring with ALOT of values:
>>>> http://www.nabble.com/file/p19338704/constantValues-many-contours.png
>>>>
>>>> Can someone explain why the square area(bottom left of the elliptical
>>>> dataset) and triangle area (bottom right) don't produce any contours? The
>>>> only similarities are that the values in those areas are constant in those
>>>> areas.
>>>> Gerrick
>>>>
>>>>
>>>> --
>>>> View this message in context:
>>>> http://www.nabble.com/Contouring-through-constant-value-areas-of-dataset-tp19
>>>> 338704p19338704.html
>>>> Sent from the VTK - Users mailing list archive at Nabble.com.
>>>>
>>>> _______________________________________________
>>>> This is the private VTK discussion list.
>>>> Please keep messages on-topic. Check the FAQ at:
>>>> http://www.vtk.org/Wiki/VTK_FAQ
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.vtk.org/mailman/listinfo/vtkusers
>>>>
>>>>         
>>     
> _______________________________________________
> This is the private VTK discussion list.
> Please keep messages on-topic. Check the FAQ at: http://www.vtk.org/Wiki/VTK_FAQ
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers
>   


-- 
Burlen Loring
Kitware, Inc.
R&D Engineer
28 Corporate Drive
Clifton Park, NY 12065-8662
Phone: 518-371-3971 x137




More information about the vtkusers mailing list