Package Summary

Tags No category tags.
Version 1.1.9
License BSD
Build type CATKIN
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/GT-RAIL/rail_pick_and_place.git
VCS Type git
VCS Version develop
Last Updated 2016-05-20
Dev Status MAINTAINED
Released RELEASED

Package Description

Construction and Use of a Recognition Database for Grasping Purposes

Additional Links

Maintainers

  • David Kent
  • Russell Toris

Authors

  • David Kent
  • Russell Toris
  • bhetherman
README
No README found. See repository README.
CHANGELOG

Changelog for package rail_recognition

1.1.3 (2015-04-14)

  • uses PCL models for generation
  • cleanup of new stuff
  • Recognition first performs a color check
  • Recognition tuning
  • added avg functions for color
  • removed metric pre-compute
  • stores metrics
  • Contributors: David Kent, Russell Toris

1.1.9 (2016-02-23)

  • Removed opencv_nonfree features
  • Update package.xml
  • Updates to suppor the Kinect2, switched to more minimal object set for demonstration purposes
  • Contributors: David Kent

1.1.8 (2015-08-18)

  • Merge branch \'develop\' of github.com:WPI-RAIL/rail_pick_and_place into develop
  • 2D image recognition added
  • Contributors: David Kent

1.1.7 (2015-05-08)

  • Calculated new orientation of combined recognized objects
  • Contributors: David Kent

1.1.6 (2015-05-05)

  • added cleared flag
  • Contributors: Russell Toris

1.1.5 (2015-04-29)

  • Bugfix for running recognition on empty object lists
  • Contributors: David Kent

1.1.4 (2015-04-27)

  • weighted average of colors
  • added doxygen
  • bug fixes in point cloud merging
  • Added remove object functionality for the recognized object list
  • Combined close together objects of the same type into single objects
  • Contributors: David Kent, Russell Toris
  • changelog updated
  • uses PCL models for generation
  • cleanup of new stuff
  • Recognition first performs a color check
  • Recognition tuning
  • added avg functions for color
  • removed metric pre-compute
  • stores metrics
  • Contributors: David Kent, Russell Toris

1.1.2 (2015-04-10)

  • removed the swap check from recognition
  • Merge branch \'develop\' of github.com:WPI-RAIL/rail_pick_and_place into develop
  • New decision tree for object models
  • docs finished
  • more docs
  • new metric trainier
  • Metric trainer filters point clouds first
  • bug removed
  • metric trainer addition
  • Added metric data collection for decision tree training
  • random selection added
  • fixes segfaults in graph search
  • graph search used
  • Contributors: David Kent, Russell Toris

1.1.1 (2015-04-06)

1.1.0 (2015-04-03)

  • new launch files
  • grasp model retriever added
  • general cleanup
  • refactor of model generator
  • major refactor of model generation
  • adds vector for search
  • more refactoring
  • begin refactor of metrics
  • Contributors: Russell Toris

1.0.4 (2015-03-31)

  • orientation added
  • Contributors: Russell Toris

1.0.3 (2015-03-31)

  • debug is now latched
  • checks for empty point cloud
  • starting to cleanup model generator
  • fix in launch
  • refactor of point cloud recognizer
  • added select all to database
  • repub of recognized objects addedg
  • Contributors: Russell Toris

1.0.2 (2015-03-30)

1.0.1 (2015-03-27)

  • Removed some old code
  • Contributors: David Kent

1.0.0 (2015-03-27)

  • object recognition listener now checks if segmentation passes in updated information vs. new object information
  • Moved actions from recognition to msgs
  • Removed unused messages
  • Message generation dependency
  • Removed some unused imports
  • Added a recognition listener that allows for shared recognition of a list of detected objects
  • Removed some unused stuff from recognition
  • Updated rail recognition to use the database, fixed a point cloud transform bug in grasp collection
  • Updates to rviz panels
  • Switched registration to use the graspdb, fixed point cloud selection in grasp collector
  • Added a vision panel that can handle segmentation and recognition, minor documentation updates to other panels
  • Moved rviz panels to rail_pick_and_place_tools, added an rviz panel for grasp collection
  • Updated grasp requests to use stamped poses instead of base_footprint frame poses
  • Switched model generation from a service to an action, and updated the rviz plugin so that it does not freeze during the model generation call
  • Updated to reflect moving some messages from rail_segmentation to rail_manipulation_messages
  • rviz plugin launch/install, new models, and some general cleanup
  • Rviz plugin updates
  • Initial rviz plugin for model generation
  • Contributors: David Kent

0.0.2 (2015-02-18)

  • catkin cleanup
  • Updated metapackage
  • Grasp collection, model building, basic recognition and grasping
  • Contributors: David Kent, Russell Toris

0.0.1 (2014-10-22)

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

  • launch/model_generator.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Version: April 1, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]
      • debug [default: false]
  • launch/metric_trainer.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Version: April 7, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]
  • launch/object_recognizer.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Version: March 31, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]
  • launch/object_recognition_listener.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Author: David Kent, Worcester Polytechnic Institute Version: March 31, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]
      • segmented_objects_topic [default: /segmentation/segmented_objects]
      • debug [default: false]
      • use_image_recognition [default: true]
  • launch/collect_images.launch
      • new_images_dir [default: $(find rail_recognition)/data/newimages/]
  • launch/rail_grasp_model_retriever.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Version: April 3, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged rail_recognition at answers.ros.org

Package Summary

Tags No category tags.
Version 1.1.9
License BSD
Build type CATKIN
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/GT-RAIL/rail_pick_and_place.git
VCS Type git
VCS Version develop
Last Updated 2016-05-20
Dev Status MAINTAINED
Released RELEASED

Package Description

Construction and Use of a Recognition Database for Grasping Purposes

Additional Links

Maintainers

  • David Kent
  • Russell Toris

Authors

  • David Kent
  • Russell Toris
  • bhetherman
README
No README found. See repository README.
CHANGELOG

Changelog for package rail_recognition

1.1.3 (2015-04-14)

  • uses PCL models for generation
  • cleanup of new stuff
  • Recognition first performs a color check
  • Recognition tuning
  • added avg functions for color
  • removed metric pre-compute
  • stores metrics
  • Contributors: David Kent, Russell Toris

1.1.9 (2016-02-23)

  • Removed opencv_nonfree features
  • Update package.xml
  • Updates to suppor the Kinect2, switched to more minimal object set for demonstration purposes
  • Contributors: David Kent

1.1.8 (2015-08-18)

  • Merge branch \'develop\' of github.com:WPI-RAIL/rail_pick_and_place into develop
  • 2D image recognition added
  • Contributors: David Kent

1.1.7 (2015-05-08)

  • Calculated new orientation of combined recognized objects
  • Contributors: David Kent

1.1.6 (2015-05-05)

  • added cleared flag
  • Contributors: Russell Toris

1.1.5 (2015-04-29)

  • Bugfix for running recognition on empty object lists
  • Contributors: David Kent

1.1.4 (2015-04-27)

  • weighted average of colors
  • added doxygen
  • bug fixes in point cloud merging
  • Added remove object functionality for the recognized object list
  • Combined close together objects of the same type into single objects
  • Contributors: David Kent, Russell Toris
  • changelog updated
  • uses PCL models for generation
  • cleanup of new stuff
  • Recognition first performs a color check
  • Recognition tuning
  • added avg functions for color
  • removed metric pre-compute
  • stores metrics
  • Contributors: David Kent, Russell Toris

1.1.2 (2015-04-10)

  • removed the swap check from recognition
  • Merge branch \'develop\' of github.com:WPI-RAIL/rail_pick_and_place into develop
  • New decision tree for object models
  • docs finished
  • more docs
  • new metric trainier
  • Metric trainer filters point clouds first
  • bug removed
  • metric trainer addition
  • Added metric data collection for decision tree training
  • random selection added
  • fixes segfaults in graph search
  • graph search used
  • Contributors: David Kent, Russell Toris

1.1.1 (2015-04-06)

1.1.0 (2015-04-03)

  • new launch files
  • grasp model retriever added
  • general cleanup
  • refactor of model generator
  • major refactor of model generation
  • adds vector for search
  • more refactoring
  • begin refactor of metrics
  • Contributors: Russell Toris

1.0.4 (2015-03-31)

  • orientation added
  • Contributors: Russell Toris

1.0.3 (2015-03-31)

  • debug is now latched
  • checks for empty point cloud
  • starting to cleanup model generator
  • fix in launch
  • refactor of point cloud recognizer
  • added select all to database
  • repub of recognized objects addedg
  • Contributors: Russell Toris

1.0.2 (2015-03-30)

1.0.1 (2015-03-27)

  • Removed some old code
  • Contributors: David Kent

1.0.0 (2015-03-27)

  • object recognition listener now checks if segmentation passes in updated information vs. new object information
  • Moved actions from recognition to msgs
  • Removed unused messages
  • Message generation dependency
  • Removed some unused imports
  • Added a recognition listener that allows for shared recognition of a list of detected objects
  • Removed some unused stuff from recognition
  • Updated rail recognition to use the database, fixed a point cloud transform bug in grasp collection
  • Updates to rviz panels
  • Switched registration to use the graspdb, fixed point cloud selection in grasp collector
  • Added a vision panel that can handle segmentation and recognition, minor documentation updates to other panels
  • Moved rviz panels to rail_pick_and_place_tools, added an rviz panel for grasp collection
  • Updated grasp requests to use stamped poses instead of base_footprint frame poses
  • Switched model generation from a service to an action, and updated the rviz plugin so that it does not freeze during the model generation call
  • Updated to reflect moving some messages from rail_segmentation to rail_manipulation_messages
  • rviz plugin launch/install, new models, and some general cleanup
  • Rviz plugin updates
  • Initial rviz plugin for model generation
  • Contributors: David Kent

0.0.2 (2015-02-18)

  • catkin cleanup
  • Updated metapackage
  • Grasp collection, model building, basic recognition and grasping
  • Contributors: David Kent, Russell Toris

0.0.1 (2014-10-22)

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

  • launch/model_generator.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Version: April 1, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]
      • debug [default: false]
  • launch/metric_trainer.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Version: April 7, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]
  • launch/object_recognizer.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Version: March 31, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]
  • launch/object_recognition_listener.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Author: David Kent, Worcester Polytechnic Institute Version: March 31, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]
      • segmented_objects_topic [default: /segmentation/segmented_objects]
      • debug [default: false]
      • use_image_recognition [default: true]
  • launch/collect_images.launch
      • new_images_dir [default: $(find rail_recognition)/data/newimages/]
  • launch/rail_grasp_model_retriever.launch
    • Author: Russell Toris, Worcester Polytechnic Institute Version: April 3, 2015
      • host [default: 127.0.0.1]
      • port [default: 5432]
      • user [default: ros]
      • password [default: ]
      • db [default: graspdb]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged rail_recognition at answers.ros.org