openmvg / awesome_3dreconstruction_list

A curated list of papers & resources linked to 3D reconstruction from images.

OpenMVG (open Multiple View Geometry)

License	Documentation	Continuous Integration (Linux/MacOs/Windows)	Build	Code Quality	Chat
	Wiki		local/docker build tutorial

Our Mission

• Extend awareness of the power of 3D reconstruction from images/photogrammetry by developing a C++ framework.

Our Vision

• Simplify reproducible research with easy-to-read and accurate implementation of state of the art and "classic" algorithms.

Our Credo

• "Keep it simple, keep it maintainable".
  • OpenMVG is designed to be easy to read, learn, modify and use.
  • Thanks to its strict test-driven development and samples, the library allows to build trusted larger systems.

Our codebase and pipeline

OpenMVG provides an end-to-end 3D reconstruction from images framework compounded of libraries, binaries, and pipelines.

• The libraries provide easy access to features like: images manipulation, features description and matching, feature tracking, camera models, multiple-view-geometry, robust-estimation, structure-from-motion algorithms, ...
• The binaries solve unit tasks that a pipeline could require: scene initialization, feature detection & matching and structure-from-motion reconstruction, export the reconstructed scene to others Multiple-View-Stereovision framework to compute dense point clouds or textured meshes.
• The pipelines are created by chaining various binaries to compute image matching relation, solve the Structure from Motion problem (reconstruction, triangulation, localization) and ...

OpenMVG is developed in C++ and runs on Android, iOS, Linux, macOS, and Windows.

Tutorials

• See Wiki
  • OpenMVG Data Structure
  • Using OpenMVG as a library
  • Using OpenMVG on your image dataset
  • ...

More information

• Authors
• Contact
• Citations
• Acknowledgements

Authors

See Authors text file

Contact

openmvg-team[AT]googlegroups.com

Citations

We are recommending citing OpenMVG if you are using the whole library or the adequate paper if you use only a submodule AContrario Ransac [3], AContrario SfM [1], GlobalSfM [4] or Tracks [2]:

@inproceedings{moulon2016openmvg,
  title={Open{MVG}: Open multiple view geometry},
  author={Moulon, Pierre and Monasse, Pascal and Perrot, Romuald and Marlet, Renaud},
  booktitle={International Workshop on Reproducible Research in Pattern Recognition},
  pages={60--74},
  year={2016},
  organization={Springer}
}

[1] Moulon Pierre, Monasse Pascal and Marlet Renaud. ACCV 2012. Adaptive Structure from Motion with a contrario model estimation.

@inproceedings{Moulon2012,
  doi = {10.1007/978-3-642-37447-0_20},
  year  = {2012},
  publisher = {Springer Berlin Heidelberg},
  pages = {257--270},
  author = {Pierre Moulon and Pascal Monasse and Renaud Marlet},
  title = {Adaptive Structure from Motion with a~Contrario Model Estimation},
  booktitle = {Proceedings of the Asian Computer Vision Conference (ACCV 2012)}
}

[2] Moulon Pierre and Monasse Pascal. CVMP 2012. Unordered feature tracking made fast and easy.

@inproceedings{moulon2012unordered,
  title={Unordered feature tracking made fast and easy},
  author={Moulon, Pierre and Monasse, Pascal},
  booktitle={CVMP 2012},
  pages={1},
  year={2012}
}

[3] Moisan Lionel, Moulon Pierre and Monasse Pascal. IPOL 2012. Automatic Homographic Registration of a Pair of Images, with A Contrario Elimination of Outliers.

@article{moisan2012automatic,
  title={Automatic homographic registration of a pair of images, with a contrario elimination of outliers},
  author={Moisan, Lionel and Moulon, Pierre and Monasse, Pascal},
  journal={Image Processing On Line},
  volume={2},
  pages={56--73},
  year={2012}
}

[4] Moulon Pierre, Monasse Pascal, and Marlet Renaud. ICCV 2013. Global Fusion of Relative Motions for Robust, Accurate and Scalable Structure from Motion.

@inproceedings{moulon2013global,
  title={Global fusion of relative motions for robust, accurate and scalable structure from motion},
  author={Moulon, Pierre and Monasse, Pascal and Marlet, Renaud},
  booktitle={Proceedings of the IEEE International Conference on Computer Vision},
  pages={3248--3255},
  year={2013}
}

Acknowledgements

openMVG authors would like to thanks libmv authors for providing an inspiring base to design openMVG. Authors also would like to thanks Mikros Image and LIGM-Imagine laboratory for support and authorization to make this library an opensource project.

awesome_3dreconstruction_list's Issues

Patented algorithms column in OpenSource SfM list?

I think it would make sense to include information which software support evading the patented feature detection and matching algorithms, either by offering a choice of alternative algorithms (actually, none of the six does that, if I'm not wrong; OpenCV does!) or accepting pre-computed data as an additional input (colmap and OpenMVG seem to be capable of that).

Update link

Link (http://cvlabwww.epfl.ch/data/multiview/) for data sets 'Dense MVS' is invalid. The data sets can be found at https://icwww.epfl.ch/~marquez/multiview/denseMVS.html now.

Add a new open source software for MVS

Hi Pierre,

Could you add this one to the list : https://github.com/flanggut/smvs

Thanks,

Stephane-lb

( There is also MVIR, but I don't think the source have been released yet )

Typo

Hi

I found a small typo in smvs license description :
BSD-3-Clquse license -> BSD-3-Clause license

Use of these algorithm

Hi Pierre,

First of all, thanks for putting this list up, it gives a good summary of what is going on in this field.
I am a physicist working for cultural heritage and I got interested in the use of 3D reconstruction on the side of my "job/phd".

However, in my field, there is a trend of "over trust" toward softwares and lack of assessment. There is not much work about accurate the measurement is (in cultural heritage I mean, not in general).
Most of them are using agisoft photoscan, which does not seem to get the best result.

On the other hand, I see that several work has been done, in particular looking at cultural heritage (ie: rebuilding Rome in one day or blunder) and many example of the algorithm here are using cultural heritage asset.

But these work are rarely reaching people in my field.
I think one of the reason is in term of access. Therefore, I was wondering how hard it is to turn these algorithm into a "basic software" like VisualSFM for example.
I saw that you are behind many executable such as openMSV or blunder, so you might be the best to ask!

Is there any technical difficulties behind or is it just matter of time/interest?
I know it is very easy to exploit command line in a GUI and exploit the executable you complied for example but I would not be that confident regarding the clean compilation and dependencies.

Additionally, how the available softwares like photoscan or realitycapture are performing compare to "the state of the art" develop in research mentioned here?

I would be very interested to push some of these tools to have a more reliable approach to 3D in heritage.

Thanks!

Please address the most recent pull request

Shall we include FSSR?

Hi, Dr. Moulon

Thank you so much for this AWESOME list! I learnt a lot from it!

I was wondering if the FSSR paper from MVE lib should be included in the surface reconstruction section as well?
Simon Fuhrmann and Michael Goesele, Floating Scale Surface Reconstruction, In: ACM Transactions on Graphics (Proceedings of ACM SIGGRAPH 2014), Vancouver, Canada, 2014.

Thank you!
John

update link for point cloud generation

The old link for "Pixelwise View Selection for Unstructured Multi-View Stereo" seems not to be working.

https://demuc.de/papers/schoenberger2016mvs.pdf, seems to work.

All in one tutorial by Yu Huang no longer available

The slides of the tutorial Introduction of Visual SLAM, Structure from Motion and Multiple View Stereo. Yu Huang 2014 are no longer available

CNN-SLAM

Hi
I recently found an amazing paper on SLAM
CNN-SLAM - http://campar.in.tum.de/Chair/ProjectCNNSLAM

Scope of the repository

I was wondering if the repo is restricted to work dealing with monocular / multiview stereo or if it accepts works based on stereo / structured devices

COLMAP is now BSD 3-clause

since version 3.5:

• https://colmap.github.io/license.html
• https://colmap.github.io/changelog.html