Are you simply looking at getting a 1-1 correspondence between the 3D stereo point cloud/2D disparity image and your 
monochrome 2D image (left/right) on the same camera? If so, this is already there for you. Usually the coordinate system 
is chosen to be the left camera, which means that all pixels in your 2D image on the left correspond to the 3D values in 
your generated point cloud.


Cheers,
Radu.


On 10/08/2010 02:34 AM, Bera, D. wrote:
> Dear all,
>
> At this moment, we are working with a standard Videre stereo camera and make use of standard nodes available from ROS repositories to publish a dense point cloud (stereo_image_proc: node).  It would be of great help if we could have some suggestions on how to correlate a point on a 2D stereo images to a 3D point in a point cloud? Do we need to make a new node to perform this operation? What would be its feasibility, accuracy?
>
> Thanks and Regards,
> Debjyoti Bera.
> _______________________________________________
> ros-users mailing list
> ros-users@code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users