<> <> == Overview == This package is an ROS wrapper for [[http://www.hitl.washington.edu/artoolkit/|ARToolkit]], which improves marker localization using point cloud data from a Kinect. [[ar_kinect]] has a single node that can be run, which takes RGB point clouds from the Kinect and outputs a transform between the camera and a recognized marker. This is based on the `ar_multi` node from the [[ar_pose]] package. The node works by creating an RGB image from the point cloud, finding the corners of the marker, and then computing a transformation to an ideal marker geometry using methods from PCL. == Nodes == {{{ #!clearsilver CS/NodeAPI node.0 { name = ar_kinect desc = This node tracks multiple AR Markers in an image/point-cloud and displays theirs relative pose. sub { 2.name = points 2.type = sensor_msgs/PointCloud2 2.desc = The point cloud data which is used for marker detection and localization. In earlier branches, this was directly mapped to /camera/rgb/points. } pub { 0.name = ar_pose_markers 0.type = ar_pose/ARMarkers 0.desc = An array of markers, each marker describe the current pose relative to the camera. 1.name = visualization_marker 1.type = visualization_msgs/Marker 1.desc = Visualization marker for [[rviz]] } param { 0.name = marker_pattern_list 0.type = string 0.desc = ARToolKit file describing each pattern. 0.default = `"$(find ar_kinect)/data/object_kinect"` 1.name = marker_data_directory 1.type = string 1.desc = Location of ARToolKit files describing each pattern. 1.default = `"$(find ar_pose)"` 2.name = threshold 2.type = double 2.desc = Threshold passed to ARToolKit 2.default = 100 3.name = publish_visual_markers 3.type = bool 3.desc = Option to enable publishing visual markers for rviz 3.default = true 4.name = publish_tf 4.type = bool 4.desc = Enable broadcasting transforms 4.default = true } prov_tf{ 0{ from = camera_frame_id to = marker_frames desc = the pose of markers relative to the camera frame } } } }}} ## AUTOGENERATED DON'T DELETE ## CategoryPackage