COLMAP

Sparse reconstruction of central Rome.

Sparse model of central Rome using 21K photos produced by COLMAP’s SfMpipeline.

Dense reconstruction of landmarks.

Dense models of several landmarks produced by COLMAP’s MVS pipeline.

About

COLMAP is a general-purpose Structure-from-Motion (SfM) and Multi-View Stereo(MVS) pipeline with a graphical and command-line interface. It offers a widerange of features for reconstruction of ordered and unordered image collections.The software is licensed under the new BSD license. If you use this project foryour research, please cite:

@inproceedings{schoenberger2016sfm,author={Sch\"{o}nberger, Johannes Lutz and Frahm, Jan-Michael},title={Structure-from-MotionRevisited},booktitle={ConferenceonComputerVisionandPatternRecognition(CVPR)},year={2016},}@inproceedings{schoenberger2016mvs,author={Sch\"{o}nberger, Johannes Lutz and Zheng, Enliang and Pollefeys, Marc and Frahm, Jan-Michael},title={PixelwiseViewSelectionforUnstructuredMulti-ViewStereo},booktitle={EuropeanConferenceonComputerVision(ECCV)},year={2016},}

If you use the image retrieval / vocabulary tree engine, please also cite:

@inproceedings{schoenberger2016vote,author={Sch\"{o}nberger, Johannes Lutz and Price, True and Sattler, Torsten and Frahm, Jan-Michael and Pollefeys, Marc},title={AVote-and-VerifyStrategyforFastSpatialVerificationinImageRetrieval},booktitle={AsianConferenceonComputerVision(ACCV)},year={2016},}

The latest source code is available atGitHub. COLMAP builds on top of existing works andwhen using specific algorithms within COLMAP, please also cite the originalauthors, as specified in the source code.

Download

Executables and other resources can be downloaded fromhttps://demuc.de/colmap/.

Getting Started

  1. Download thepre-built binaries or build thelibrary manually fromsource(seeInstallation).

  2. Download one of the provided datasets (seeDatasets)or use your own images.

  3. Use theautomatic reconstruction to easily build modelswith a single click (seeQuickstart).

Support

Please, useGitHub Discussionsfor questions and theGitHub issue trackerfor bug reports, feature requests/additions, etc.

Acknowledgments

COLMAP was originally written byJohannes Schönberger withfunding provided by his PhD advisors Jan-Michael Frahm and Marc Pollefeys.The team of core project maintainers currently includesJohannes Schönberger,Paul-Edouard Sarlin, andShaohui Liu.

The Python bindings in PyCOLMAP were originally added byMihai Dusmanu,Philipp Lindenberger, andPaul-Edouard Sarlin.

The project has also benefitted from countless community contributions, includingbug fixes, improvements, new features, third-party tooling, and communitysupport (special credits toTorsten Sattler).