ubuntu 12.10 + groovy + openni + USB3.0 = working?

edit

Let me share what I found out:

1. I performed a fresh installation of Ubuntu 12.10 and ROS Groovy following the standard procedures.

2. Installed the openni packages:

   sudo apt-get install ros-groovy-openni-*

3. Plugged a Kinect device on a USB2.0 port

4. Launched openni:

   roslaunch openni_launch openni.launch

5. openni_launch dies and complains about rosdep, recommends running: "sudo rosdep init" and "rosdep update". I do so.

6. re-run "roslaunch openni_launch openni.launch". This time it seems to be SUCCESSFUL.

You would think that everything works, but we are not so lucky:

1. Did rosrun rviz rviz. Try to visualize the topic /camera/depth_registered/points: Nothing appears. FAIL
2. Did rostopic hz /camera/depth_registered/points: No response, apparently there is no messages being published on that topic. EPIC FAIL
3. Did rosrun image_view image_view image:=/camera/rgb/image_raw: The image gets displayed, but in grayscale (WTF??). Well, at least there is communication with the device. PARTIAL SUCCESS
4. Did rosrun rviz rviz. Try to visualize the topic /camera/depth/points: It works! SUCCESS

It looks like there is some kind of trouble with the RGB image and that is causing the /camera/depth_registered/points not being published.

Next I am going to test the device connected to a USB3.0 port. I will update the answer with the new findings.

[UPDATE]

Here is the new information:

1. Connected a Kinect to USB3.0 port.
2. Did roslaunch openni_launch openni.launch: Launches normally. SUCCESS
3. Did rosrun rviz rviz. Try to visualize the topic /camera/depth_registered/points: Same issue as before. FAIL
4. Did rosrun image_view image_view image:=/camera/rgb/image_raw: The image gets displayed, but in grayscale. Just like before. PARTIAL SUCCESS
5. Did rosrun rviz rviz. Try to visualize the topic /camera/depth/points: It works. SUCCESS

So to sum up and answer this question:

Ubuntu 12.10 + groovy + openni + USB3.0 = partially works (there is issues with the color image and the color pointcloud)