[IGSTK-Users] Augmented reality

Alvaro Gómez agomez at fing.edu.uy
Mon Jun 1 15:19:19 EDT 2009


Hi:

For an augmented reality application we need to mix video and some
spatial objects. We are taking the following approach that we want to
validate with you:

a) A modified View class with more than one renderer.

a.1) We are adding a second renderer(background_renderer) to the View.
The background_renderer is used to render the video.
The m_Renderer and the background_renderer are put in two different
layers of the m_RenderWindow (layers 1 and 0 respectively)

a.2) The View class has a method to add an object representation to the
background_renderer in the same way as the AddObjectProcessing does with
the m_Renderer

a.3) The RefreshRender is modified to iterate and request the update
also on the background objects (the objects added to the
background_renderer with a.2)

b) For the video we are using Imager-ImagerTool and
VideoFrameSpatialObject-VideoFrameRepresentation as in the
VideoFrameGrabberAndViewerWebcamWin example.

c) In the AR application, the VideoFrameRepresentation is added to the
background renderer of the View.

d) The video is then rendered using a parallel projection (camera of the
background_renderer) and the objects in the foreground are rendered with
  a camera (camera of the foreground m_Renderer) with parameters that
match the real camera.


Probably there is a better way to do the same thing or a way without
having to modify the View class. Please let us know your opinion.

Thanks.
Alvaro




More information about the IGSTK-Users mailing list