Difference between two versions of openni_launch
Hi,
I am using the Kinect sensor with the openni driver and the Roboearth software for object modeling and detection. While going through all the tutorials and setting this up, I found out that there are two (maybe more?) ways to call openni_launch. Each version seems to publish the same information but have different topics.
The openni wiki (http://www.ros.org/wiki/openni_launch) launches openni using: roslaunch openni_launch openni.launch. Amongst many others, this version publishes a topic /camera/depth_registered. Also, the fixed frame for rviz is set to /camera_depth_optical_frame (im guessing we could also set it to /camera_rgb_optical_frame). This version also does not publish a topic /camera/rgb/points.
In the roboearth tutorail (http://www.ros.org/wiki/roboearth/Tutorials/Record%20an%20object%20model) we launch openni using roslaunch openni_camera openni_node.launch. As mentioned this provides the same information but publishes slightly different topics. In particular, using this method we have to set rviz fixed frame to /opnni_rgb_optical_frame. Also, this publishes a topic called /camera/rgb/points which is used by the software.
My question, what is the difference (if any) between these two ways of launch the openni driver? Is one version better than another? Does it matter?
Currently, I am using the 2nd version to launch the driver as it publishes to topics that are subscribed by the roboearth software. I am using ros_electric on Ubuntu 11.04 64-bit.
Thanks!
