[vtkusers] Changing the projection matrix

[Cameron Lowell Palmer]

Maybe some more explanation is necessary? 

What I have:
* Video camera calibration with intrinsics of the camera (i.e. focal length, center).
* Framemarker tracking that returns a rotation/translation matrix that represent the marker in space.
* VES/VTK app that loads the image plane and several other objects.
* pose matrix, camera calibration, and streaming video frames are all available to VES/VTK

The specific experiment I’m trying to complete is placing a logo over the framemarker in the image.

When doing this with raw OpenGL I need to set the OpenGL projection matrix so that it conforms to the camera intrinsics. 
Then you draw the objects in the scene with the pose matrix loaded. However, translating this to VES/VTK is defeating me.

Problems I think I have:
* The camera image plane is now in 3-space.
* I can’t directly apply the projection matrix to vesCamera. However, it seems I can set Focal Point and View Angle.
* I need to translate objects into the coordinate space of camera image plane

So, am I correctly identifying my problems? Are there problems I missed? Has anyone developed a working example of an augmented reality app using VES/VTK or even just VTK?

Thanks,

cameron.



> On 21. jan. 2015, at 12.32, Cameron Lowell Palmer <cameron.palmer at ntnu.no> wrote:
> 
> I’m trying to use VES/VTK in support of an Augmented Reality app on iOS and I need to provide a new projection matrix that I generated using the Camera Calibration. Is this not supported or do I have to completely bypass VTK and handle it in the shader?