[vtkusers] Question about surface normal

[Luca Pamparana]

Hi David,

Thanks for your reply. So, if I have got it correct:

- I can take vectors AC and BD and using the cross product find the
current normal vector.
- The normal vector I want would be described as user_specified_point
- (any of my polygon points). Is this correct?
- Then I could use your transformation function to transform the 4 points...

If I was to use a vtkPlane object, would this be easier? In the
example that you send me, the function specifies the center of the
plane and the surface normal. Would the plane center be invariant
under this transformation? I would have thought that if I was to
specify a normal explicitly, the points would have to be
recalculated....

Sorry for the constant mailing. I hope this would be the final question!

Thanks,

Luca

On Thu, Nov 5, 2009 at 2:44 PM, David Doria <daviddoria+vtk at gmail.com> wrote:
> On Thu, Nov 5, 2009 at 9:40 AM, Luca Pamparana <luca.pamparana at gmail.com> wrote:
>> Hello David,
>>
>> Thanks for your reply and the tip about using the smart pointers!
>>
>> I was actually not sure how to generate the normal vector that would
>> point to a user specified point. So, my plane has 1 cell and 4 points.
>> I have one user specified point say (0, 0, 0). What would be a valid
>> unit normal vector that will be normal to the whole plane? I know that
>> the dot product of two normal vectors is 0 but I am struggling to
>> figure out how to use this in this case.
>>
>> Sorry this is a real noob question but would be really obliged for any
>> help here...
>>
>> Thanks,
>>
>> Luca
>
>
> Say your polygon is composed of points A, B, C, D.
>
> The cross product of any two vectors between any of those points will
> give you a vector normal to the polygon (we are assuming here that
> A,B,C and D actually lie on a plane).
>
> You can then use the function I sent to align this normal with the
> vector you want to be the new normal. Apply the transformation that is
> computed in that function to all of your points and you should end up
> with a polygon with the normal you want. You will have to align the
> center of mass of the points with the origin before applying the
> rotation though so that you get the expected result.
>
> Thanks,
>
> David
>