ifco_pose_estimator

ROS package that estimates the pose of an IFCO with respect to a camera.

Installation

• Installing the following packages via apt should be enough for building ifco_pose_estimator:
  • ros-ROS_DISTRO-ros-core
  • ros-ROS_DISTRO-pcl-ros
  • ros-ROS_DISTRO-moveit
• ifco_pose_estimator also uses the OpenCV, Boost and PCL libraries, which will be installed by installing the packages above.
• Clone this repository into a catkin workspace and run catkin_make.
• The package has been tested on Ubuntu 14.04 with ROS Indigo and Ubuntu 16.04 with ROS Kinetic.

Functionality

The package performs the following steps:

• The ifco_pose_server subscribes to the pcl_topic and keeps only the points of the point cloud that might correspond to the IFCO based on thresholding of the A channel of an image of the scene in LAB format.
• The ifco_pose_server fits an IFCO model to the resultant point cloud using ICP and the ifco_pose service returns the estimated pose of the IFCO with respect to the camera and the achieved fitness. The IFCO coordinate frame is placed at the centre of the IFCO base with the z axis pointing away from the opening of the IFCO and the x axis pointing along the long side of the IFCO.
• If the ifco_pose service was called with the parameter publish_ifco set to True, then the ifco_pose_server will also publish the IFCO model in the planning scene.

Node input

The user must provide the node with the following information (usually via launch_ifco_pose_server.launch ):

• A pcl_topic to subscribe to.
• Parameters A_high and A_low with which the aforementioned thresholding will be performed.
• Maximum number of icp_iterations that ICP will perform to converge.
• An initial_pose_estimate of the pose of the IFCO with respect to the camera.
• Parameter use_initial_pose_estimate to denote whether the user's initial_pose_estimate will be used.

Notes

• All topics and tfs referenced above are expressed wrt the camera frame. This is the camera_rgb_optical_frame for the Primesense or the kinect_ir_frame for the Kinect.
• If use_initial_pose_estimate is set to 0, then the node will try to compute an estimate of the IFCO pose by averaging the position of the points of the point cloud that might correspond to the IFCO and rotating the model by -45 degrees along the x axis of the IFCO (this is based on the assumption that the camera will be usually overlooking the IFCO).

How to run

• Create a node that launches your camera and maps the proper frame to the camera frame.
• Point the camera to an IFCO and do roslaunch ifco_pose_estimator launch_ifco_pose_server.launch
• Open Rviz.
• Do rosrun ifco_pose_estimator ifco_pose_client to test the node.