I am currently working on 3d LIDAR SLAM and trying to apply LOAM, LIO-SAM etc. I have already read some of these papers to get a better idea of the way they work. However, i did not find anything related to loop closure in these algorithms as opposed to visual(camera-based) SLAM.
I am wondering whether loop closure is not often used in lidar slam(if so, then how to get rid of drift) or they just use some additional camera which does bag of words based loop closure(as in orb-slam for example).