search

tum-vision / lsd_slam

LSD-SLAM
GNU General Public License v3.0
2.62k stars 1.23k forks source link

Which is the best SLAM algorithm? #246

Open apesIITM opened 7 years ago

apesIITM commented 7 years ago

Hello, I have tried running LSD-SLAM and RTAB-SLAM, and currently reading about few more papers. Can you point out in what ways is LSD-SLAM better. (As per my experiments, RTAB-SLAM gave better map representation) [ No offence intended to anyone :) ] Thanks.

Dhivin commented 7 years ago

I am currently looking in moncular slam techniques? Can any suggest me the best monocular slam method?

proepkes commented 7 years ago

ElasticFusion is also pretty good

mihaibujanca commented 7 years ago

Elastic fusion is not monocular though

mihaibujanca commented 7 years ago

The answer is, of course, it depends. What are you trying to achieve? Techincally SLAM - Simultaneous localisation and mapping has two parts to it: The localisation - how well the system manages to localise itself relative to the space it's in - and subsequently relocalise in an already visited region.

The mapping - how accurately it can construct a map of the environment. The form this map takes is can massively vary from reconstructing the trajectory of the device (odometry map) to performing dense 3D reconstruction.