Hi there,
In the openni_camera package I have changed the frame id conventions to
conform with REP 103. [1]
I've documented it at
http://www.ros.org/wiki/openni_cameraThe frames which exist are:
/openni_camera
|
|> /openni_rgb_frame
| |
| |> /openni_rgb_optical_frame (Z forward)
|
|> /openni_depth_frame
|
|> /openni_depth_optical_frame (Z forward)
I've also added a simple launch file to publish the transforms between
these frames in lauch/openni_frames.launch The numbers of these
transforms are currently gross approximations. And when they become
accurate we will likely want to have seperate files for each device.
I've added this launch file in the openni_kinect.launch as in include as
well.
This has been done to support the pointcloud_to_laserscan package which
I am adding right now as well. This package takes the kinect's depth
cloud and simulates a laser scan sampling through the data. This will
allow us to use our standard 2d laser based navigation algorithms with
the much cheaper Kinect.
The new package is in the ni stack. And there is an example launch file:
roslaunch pointcloud_to_laserscan kinect_laser.launch
This will produce a topic "scan" in you can see the output of the
simulated laser.
Tully
[1]
http://www.ros.org/reps/rep-0103.html_______________________________________________
Ros-kinect mailing list
[hidden email]
https://code.ros.org/mailman/listinfo/ros-kinect
Locked
