GLOMAP: Global Structure-from-Motion Revisited

Project page | Paper

About

GLOMAP is a general purpose global structure-from-motion pipeline for image-based reconstruction. GLOMAP requires a COLMAP database as input and outputs a COLMAP sparse reconstruction. As compared to COLMAP, this project provides a much more efficient and scalable reconstruction process, typically 1-2 orders of magnitude faster, with on-par or superior reconstruction quality.

If you use this project for your research, please cite

@inproceedings{pan2024glomap,
    author={Pan, Linfei and Barath, Daniel and Pollefeys, Marc and Sch\"{o}nberger, Johannes Lutz},
    title={{Global Structure-from-Motion Revisited}},
    booktitle={European Conference on Computer Vision (ECCV)},
    year={2024},
}

Using pixi

Easiest way to run this example is using pixi, you can easily install it using curl explained here https://pixi.sh/latest/#installation

then

pixi run example

End-to-End with Docker.

Download one of the official datasets unzip it. For example if we download the south-building.zip

mkdir -p data
unzip south-building.zip -d data/

Build and enter the docker container by running the following scripts:

cd docker
./build.sh # Builds the docker image
./start.sh # Starts the docker container
./attach.sh # Enter the container

When we're inside the container we can build glomap and install it.

mkdir -p build
cd build
cmake .. -GNinja
ninja
ninja install

To run it we must first extract the features and match them with colmap.

cd /ws
colmap feature_extractor \
    --image_path    data/south-building/images \
    --database_path data/south-building/database.db

colmap sequential_matcher \
    --database_path data/south-building/database.db 

Now we're ready to run and visualize glomap:

glomap mapper \
    --database_path data/south-building/database.db \
    --image_path    data/south-building/images \
    --output_path   data/south-building/sparse

To cleary see what's happening it's recommended you open the blueprint.rbl file in this directory. If you want a cooler visualization you can resume a mapping, you do this by first running the steps above and then running the following command:

glomap mapper_resume \
    --image_path data/south-building/images \
    --input_path data/south-building/sparse/0 \
    --output_path data/south-building/sparse

If you're experiencing performance issues with rerun you might have to run the viewer outside of the docker container.

From video

To run glomap on a video you will first have to split it into multiple frames and put them in a directory. To easily do this you can use the the script in scripts/extract_frames.py like this:

# Split 'colosseum.mp4' into frames at 1 frame per second and puts the images in colosseum/images.
scripts/extract_frames.py -v colosseum.mp4 -o colosseum/images --desired-fps 1

# Then you can extract the features like normally.
colmap feature_extractor \
    --image_path    colosseum/images \
    --database_path colosseum/database.db
  colmap sequential_matcher \ 
    --database_path colosseum/database.db 

glomap mapper \
    --database_path colosseum/database.db \
    --image_path    colosseum/images \
    --output_path   colosseum/sparse

Notes

• For larger scale datasets, it is recommended to use sequential_matcher or vocab_tree_matcher from COLMAP.

colmap sequential_matcher --database_path DATABASE_PATH
colmap vocab_tree_matcher --database_path DATABASE_PATH --VocabTreeMatching.vocab_tree_path VOCAB_TREE_PATH

• Alternatively, one can use hloc for image retrieval and matching with learning-based descriptors.

Acknowledgement

We are highly inspired by COLMAP, PoseLib, Theia. Please consider also citing them, if using GLOMAP in your work.

Support

Please, use GitHub Discussions at https://github.com/colmap/glomap/discussions for questions and the GitHub issue tracker at https://github.com/colmap/glomap for bug reports, feature requests/additions, etc.

Contribution

Contributions (bug reports, bug fixes, improvements, etc.) are very welcome and should be submitted in the form of new issues and/or pull requests on GitHub.

License

Copyright (c) 2024, ETH Zurich.
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:

    * Redistributions of source code must retain the above copyright
      notice, this list of conditions and the following disclaimer.

    * Redistributions in binary form must reproduce the above copyright
      notice, this list of conditions and the following disclaimer in the
      documentation and/or other materials provided with the distribution.

    * Neither the name of ETH Zurich nor the names of its contributors may
      be used to endorse or promote products derived from this software
      without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.