Only released in EOL distros:  

camera_self_filter

  • camera_self_filter: The camera_self_filter node renders the mesh model of its current configuration into a camera view. It publishes a binary mask image stream and optionally the original camera image with the binary mask in the alpha channel as shown below.

    Note: The node requires an X-server to render the mask. It is recommended to run it on an external Ubuntu desktop machine with OpenGL enabled graphics. If the node is to be run on the PR2 check out pr2_camera_self_filter package.

  • camera_self_filter_server: Does the same as the camera_self_filter node but does not stream constantly, but rather creates a service that returns the binary mask of the robot configuration for the requested time_stamp.
  • image_view: A modified version of image_view to display the overlayed image.

rosrun camera_self_filter image_view image:=/wide_stereo/left/image_rect_color_alpha_masked

attachment:overlayed_image.jpg

  • The mask can be displayed running

rosrun image_view image_view image:=/wide_stereo/left/self_mask

attachment:mask.jpg

launch

Example launch file:

<node pkg="camera_self_filter" type="camera_self_filter"  
 name="camera_self_filter_wide_stereo_left" output="screen">
   <param name="camera_frame" type="string" value="wide_stereo_optical_frame" />
   <param name="camera_topic" type="string" value="/wide_stereo/left" />
   <param name="publish_mask" type="bool" value="true" />
   <param name="publish_alpha" type="bool" value="true" />
   <param name="inverted" type="bool" value="false" />
   <param name="robot_description_package_path" type="string" 
     value="$(find pr2_description)" />
                
</node>

Note that the camera_self_filter node expects the robot urdf on the parameter server. If has not been loaded automatically (for instance when replaying a .bag file), set the "robot_description" parameter manually (e.g. run "cat robot.urdf | xargs rosparam set robot_description").

Wiki: camera_self_filter (last edited 2011-05-19 17:33:54 by ChristianBersch)