maxibooksiyi/manhattanslam

A Planar RGB-D SLAM which utilizes Manhattan World structure to provide optimal camera pose trajectory while also providing a sparse reconstruction containing points, lines and planes, and a dense surfel-based reconstruction.

ManhattanSLAM

Authors: Raza Yunus, Yanyan Li and Federico Tombari

ManhattanSLAM is a real-time SLAM library for RGB-D cameras that computes the camera pose trajectory, a sparse 3D reconstruction (containing point, line and plane features) and a dense surfel-based 3D reconstruction. Further details can be found in the related publication. The code is based on ORB-SLAM2.

Related Publication:

Raza Yunus, Yanyan Li and Federico Tombari, ManhattanSLAM: Robust Planar Tracking and Mapping Leveraging Mixture of Manhattan Frames, in 2021 IEEE International Conference on Robotics and Automation (ICRA) . PDF.

1. License

ManhattanSLAM is released under a GPLv3 license. For a list of all code/library dependencies (and associated licenses), please see Dependencies.md.

If you use ManhattanSLAM in an academic work, please cite:

@inproceedings{yunus2021manhattanslam,
    author = {R. Yunus, Y. Li and F. Tombari},
    title = {ManhattanSLAM: Robust Planar Tracking and Mapping Leveraging Mixture of Manhattan Frames},
    year = {2021},
    booktitle = {2021 IEEE international conference on Robotics and automation (ICRA)},
}

2. Prerequisites

We have tested the library in Ubuntu 16.04 and Ubuntu 20.04, but it should be easy to compile on other platforms. A powerful computer (e.g. i7) will ensure real-time performance and provide more stable and accurate results. Following is the list of dependecies for ManhattanSLAM and their versions tested by us:

• OpenCV: 3.3.0, 3.4.3
• PCL: 1.7.2, 1.10
• Eigen3: 3.3
• DBoW2: Included in Thirdparty folder
• g2o: Included in Thirdparty folder
• Pangolin
• tinyply: 2.3.2

3. Building and testing

Clone the repository:

git clone https://github.com/razayunus/ManhattanSLAM

There is a script build.sh to build the Thirdparty libraries and ManhattanSLAM. Please make sure you have installed all required dependencies (see section 2). Execute:

cd ManhattanSLAM
chmod +x build.sh
./build.sh

This will create libManhattanSLAM.so in lib folder and the executable manhattan_slam in Example folder.

To test the system:

1. Download a sequence for one of the following datasets and uncompress it:

   • TUM RGB-D: https://vision.in.tum.de/data/datasets/rgbd-dataset
   • ICL-NUIM: https://www.doc.ic.ac.uk/~ahanda/VaFRIC/iclnuim.html
   • TAMU RGB-D: http://telerobot.cs.tamu.edu/MFG/rgbd/livo/data.html

2. Associate RGB images and depth images using the python script associate.py. You can generate an associations file by executing:

python associate.py PATH_TO_SEQUENCE/rgb.txt PATH_TO_SEQUENCE/depth.txt > associations.txt

Note: For ICL-NUIM sequences, the association files are already given but the association is defined as depth > rgb rather than rgb > depth. This can be changed by transforming associations.txt as:

cat associations.txt | sed 's/depth/temp/g;s/rgb/depth/g;s/temp/rgb/g' | tee associations.txt > /dev/null

3. Execute the following command. Change Config.yaml to ICL.yaml for ICL-NUIM sequences, TAMU.yaml for TAMU RGB-D sequences or TUM1.yaml, TUM2.yaml or TUM3.yaml for freiburg1, freiburg2 and freiburg3 sequences of TUM RGB-D respectively. Change PATH_TO_SEQUENCE_FOLDERto the uncompressed sequence folder. Change ASSOCIATIONS_FILE to the path to the corresponding associations file.

./Example/manhattan_slam Vocabulary/ORBvoc.txt Example/Config.yaml PATH_TO_SEQUENCE_FOLDER ASSOCIATIONS_FILE