Multi-view triangulation

Reconstruct a point from N image projections

121 Downloads

Updated 7 Aug 2020

View License

This function implements multi-view triangulation for reconstructing a 3D point from its projections in N
images and the corresponding camera matrices. It is the extension in N views of my stereoReconsPts
function (see https://www.mathworks.com/matlabcentral/fileexchange/67383-stereo-triangulation).

At the user's choice, a point can be reconstructed by any of the following multi-view triangulation methods:
i) The N view DLT
ii) The N view inhomogeneous method
iii) The N view midpoint method
vi) A N view robust method that can tolerate mismatches

Note that i) & ii) are algebraic methods, whereas iii) involves a more meaningful geometric constraint.

Since the rays defined by the camera centers and corresponding image projections will in general be skew
(i.e., they will not meet in a single intersection point), it is recommended to refine the estimate
computed above with one of the following methods:
i) Minimization of the reprojection error in N views with VGG's Gauss-Newton code
ii) Minimization of the reprojection error in N views with Levenberg-Marquardt
iii) Minimization of the N view angular criterion by Recker et al: "Statistical angular error-based
triangulation for efficient and accurate multi-view scene reconstruction", WACV 2013

In general, the combinations 'dlt', 'lm' or 'midp', 'lm' should work well in most cases.
Refer to the comments in the code for more details.

* If you use this function in your published work, please cite the following paper:

S. Nousias, M. Lourakis, C. Bergeles (2019). Large-Scale, Metric Structure from Motion for Unordered Light Fields
IEEE Conf. on Pattern Recognition and Computer Vision (CVPR'19), pp. 3292-3301.
https://openaccess.thecvf.com/content_CVPR_2019/papers/Nousias_Large-Scale_Metric_Structure_From_Motion_for_Unordered_Light_Fields_CVPR_2019_paper.pdf

@inproceedings{nousias2019large,
title={Large-Scale, Metric Structure From Motion for Unordered Light Fields},
author={S. Nousias and M. Lourakis and C. Bergeles},
booktitle={IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2019},
pages={3287--3296}
}

Cite As

Manolis Lourakis (2023). Multi-view triangulation (https://www.mathworks.com/matlabcentral/fileexchange/78955-multi-view-triangulation), MATLAB Central File Exchange. Retrieved .

MATLAB Release Compatibility
Created with R2013b
Compatible with any release
Platform Compatibility
Windows macOS Linux
Tags Add Tags
Acknowledgements

Inspired by: Stereo triangulation

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!
Version Published Release Notes
1.0.2

Changed title

1.0.1

Updated description

1.0.0