tf configuration issues.

[tanmayshankar]

Hello everyone,

The ccny_rgbd package is compiled and ready to go, (Ubuntu 12.04, ROS-Hydro). Right now I run it according to the readme, with publish cloud set to true. I'm using a ROSbag with kinect data, so I have the device nodelet disabled.

When I open up rviz and load the vo+mapping.rviz config file, the kinect frame seems to be moving around with respect to my world frame, and the other frames (openni_camera, optical_frame for rgb and depth) are at some transformation to the kinect frame. Apart from this however, there is nothing showing up in rviz. No path, no model, even after I fiddled with topics of the rviz visualization markers.

The ROS wiki says it needs the transform between camera_link and the moving camera frame.
From my understanding, this should just be a transform broadcaster publishing a fixed transform between camera_link and openni_camera OR kinect, but that doesn't seem to work.

Can anyone tell me what the correct tf configuration is? Also, What are the corresponding changes in my rviz config file, considering the tf tree the rviz file has is between /odom and /camera_link ?

Thanks!