3D reconstruction of points from three views - is there a 3 camera alternative to triangulate?

56 views (last 30 days)
Hi everybody,
I have three sets of x,y data, taken from an object tracked as it moves through three dimensions and taken by three different cameras.
I have the camera parameters and matrices, I can calibrate the images etc and have even been able to fully reconstruct a 3D point cloud based on two cameras, using the Matlab 'triangulate' function. But I was wondering if there is some Matlab function or code that can accept more than two cameras/views? I would like to be able to use the information from all three cameras...
Thanks very much for any help,

Accepted Answer

Dima Lisin
Dima Lisin on 4 Dec 2015
Edited: Dima Lisin on 16 Mar 2016
The triangulate function can only handle two views. However, the algorithm that it uses can be easily extended to multiple views. Please see Multiple View Geometry in Computer Vision by Harley and Zisserman for details.
Edit: As of release R2016a you can use triangulateMultiview to do 3D reconstruction from any number of views.
cui on 9 Apr 2021
This internal function can be reconstructed from multiple points.
worldPoints = vision.internal.triangulateMultiViewPoints(pointTracks, cameraMatrices)

Sign in to comment.

More Answers (0)


Find more on MATLAB Support Package for USB Webcams in Help Center and File Exchange

Community Treasure Hunt

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

Start Hunting!

Translated by