Transform between Kinect skeleton base frame and camera frame
I am trying to use the ROS NI drivers (openni_camera) with the Kinect, and additionally use openni_tracker to get the transforms for the joints for a detected skeleton. However, the joint transforms are published under the "/openni_depth" frame, which doesn't exist in the tree of frames published by the kinect_frames.launch file. Setting an identity static transform between /openni_depth_frame and /openni_depth didn't seem to give correct results.
Does anybody know the transform between /openni_depth and a frame in the hierarchy published by the driver?
Thanks,
Peter Brook