Examples of fulling working visual SLAM using GTSAM?

[Nicolas Perez]

I was wondering if anyone knew of a working visual SLAM system that uses GTSAM. In particular, I'm interested in using ORB features with a depth (or stereo) camera. 

I'm wondering how people usually handle outliers in the visual feature extraction process.

[Dellaert, Frank]

Yes, Davide’s group in Zurich has SVO, very performant, Based on work Christian Forster did with us, with Luca Carlone.

We have done VO in the past, but never released. With stereo actually :-)

We are also working on GTSFM in a separate repo, which is general SfM, although it doesn’t create rival COLMAP yet - very much alpha…

Best!

Frank

---

From: gtsam...@googlegroups.com <gtsam...@googlegroups.com> on behalf of Nicolas Perez <nicolasa...@gmail.com>
Sent: Thursday, November 18, 2021 5:06:38 PM
To: gtsam users <gtsam...@googlegroups.com>
Subject: [GTSAM] Examples of fulling working visual SLAM using GTSAM?

 

I was wondering if anyone knew of a working visual SLAM system that uses GTSAM. In particular, I'm interested in using ORB features with a depth (or stereo) camera. 

I'm wondering how people usually handle outliers in the visual feature extraction process.

--
You received this message because you are subscribed to the Google Groups "gtsam users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gtsam-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gtsam-users/16131da0-5710-4583-8bd1-6869ee96aaaan%40googlegroups.com.

[dav...@oxfordrobotics.institute]

A good reference example is Kimera by Luca Carlone's group: https://github.com/MIT-SPARK/Kimera 

They use GTSAM for their back-end.