About this fork

• Modified to work with custom RGB-D datasets with known Poses.
• Can run online with BenchBot System (https://github.com/qcr/benchbo.git)
• Example Script to publish data publish_data/benchbot_ros.py
• Saves maps in ${HOME}/.ros at the end.
• Saves Axis-aligned Cuboids of all semantic instances in the map automatically at the end in ${HOME}/.ros/objectmap.json
• Warning: Just hacky solution .. works for my case. !! Use at your own risk !!
• Uses depth_segmentation from this fork https://github.com/suuman/depth_segmentation_benchbot.git

Voxblox++

Voxblox++ is a framework for incrementally building volumetric object-centric maps during online scanning with a localized RGB-D camera. Besides accurately describing the geometry of the reconstructed scene, the built maps contain information about the individual object instances observed in the scene. In particular, the proposed framework retrieves the dense shape and pose of recognized semantic objects, as well as of newly discovered, previously unobserved object-like instances.

Getting started

• Installing on Ubuntu
• Sample Datasets
• Basic usage
• The Voxblox++ node

More information and sample datasets can be found in the wiki pages.

Citing

The Voxblox++ framework is described in the following publication:

• Margarita Grinvald, Fadri Furrer, Tonci Novkovic, Jen Jen Chung, Cesar Cadena, Roland Siegwart, and Juan Nieto, Volumetric Instance-Aware Semantic Mapping and 3D Object Discovery, in IEEE Robotics and Automation Letters, July 2019. [PDF] [Video]

@article{grinvald2019volumetric,
  author={M. {Grinvald} and F. {Furrer} and T. {Novkovic} and J. J. {Chung} and C. {Cadena} and R. {Siegwart} and J. {Nieto}},
  journal={IEEE Robotics and Automation Letters},
  title={{Volumetric Instance-Aware Semantic Mapping and 3D Object Discovery}},
  year={2019},
  volume={4},
  number={3},
  pages={3037-3044},
  doi={10.1109/LRA.2019.2923960},
  ISSN={2377-3766},
  month={July},
}

The original geometry-only framework was introduced in the following publication:

• Fadri Furrer, Tonci Novkovic, Marius Fehr, Abel Gawel, Margarita Grinvald, Torsten Sattler, Roland Siegwart, and Juan Nieto, Incremental Object Database: Building 3D Models from Multiple Partial Observations, in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), October 2018. [PDF] [Video]

@inproceedings{8594391,
  author={F. {Furrer} and T. {Novkovic} and M. {Fehr} and A. {Gawel} and M. {Grinvald} and T. {Sattler} and R. {Siegwart} and J. {Nieto}},
  booktitle={2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  title={{Incremental Object Database: Building 3D Models from Multiple Partial Observations}},
  year={2018},
  pages={6835-6842},
  doi={10.1109/IROS.2018.8594391},
  ISSN={2153-0866},
  month={Oct},
}

If you use Voxblox++ in your research, please cite accordingly.

License

The code is available under the BSD-3-Clause license.