Vision Tutorial¶

This doc provides steps to run MSFT Kinect with ROS¶

Assumes that ROS, openni_kinect are installed:
http://www.ros.org/wiki/electric/Installation/Ubuntu

Listen for and view Camera input
follow steps listed here: http://www.ros.org/wiki/openni_camera

Save sensor data to file (and create map, in theory)
Follow steps here: http://www.ros.org/wiki/slam_gmapping/Tutorials/MappingFromLoggedData#Building_a_map
NOTE: When "making your own bag," replace /base_scan with /camera/depth/image_raw

Note (from Nick): The data type in the bag that gmapping reads has both tf messages and laser scan messages. So we'd need to send something other than imageraw, like a bag file with topics /tf and with something converted to sensor_msgs/LaserScan (there are ways to convert the points to that, for instance), or find a way to get the kinect to output sensor_msgs/laser_scan.

Next steps:
http://www.ros.org/wiki/rgbdslam
(rgbdslam may have all the features of slam_gmapping and more, so we may not need the above step)