Only released in EOL distros:  

This package can extract semantic information from a geometric map.

ROS API

semantic_extraction_node

The semantic_extraction_node gets the geometric map via service call from cob_3d_mapping_geometry_map and performs a tabletop extraction. The tables found are published as cob_3d_mapping_msgs/ShapeArray and as visualization_msgs/Marker.

Parameters

~tilt_angle (double, default: 3.0)
  • The maximum angle deviation allowed for the table plane (in degrees)
~height_min (double, default: 0.6)
  • The minimum height for the table plane
~height_max (double, default: 1.2)
  • The maximum height for the table plane
~area_min (double, default: 0.5)
  • The minimum area for the table plane
~area_max (double, default: 3)
  • The maximum area for the table plane

Usage/Examples

Start the mapping pipeline in cob_3d_mapping_pipeline

roslaunch cob_3d_mapping_pipeline mapping.launch

Perform a 360° scan of the room

rosrun cob_3d_mapping_pipeline test_360_scan.py

Check the consistency of the point map and the geometric map in rviz.
Start the semantic extraction of tables

roslaunch cob_3d_mapping_semantics extract_tables.launch

Trigger the table detection by running

roslaunch cob_3d_mapping_semantics get_tables_client.launch

Start moving to a table

roslaunch cob_3d_mapping_semantics move_to_table.launch

Wiki: cob_3d_mapping_semantics (last edited 2012-11-07 13:31:32 by ShaghayeghNazari)