Difference between revisions of "Simultaneous localization and mapping"

From XVRWiki
Jump to navigation Jump to search
 
(14 intermediate revisions by 2 users not shown)
Line 1: Line 1:
'''Simultaneous localization and mapping''' is a method of [[3D tracking]] based on camera data of an environment. It uses one or more cameras that are mounted on a [[rigid body]] like a headset. It operates based on computer vision, heavily relying on feature detection. It is similar in concept to [[visual inertial odometry]] (VIO).
+
'''Simultaneous localization and mapping''' (SLAM) is a method of [[3D tracking]] that uses cameras to capture images of an environment and uses computer vision to determine the pose of a rigid body attached to the camera(s). It is a kind of [[egocentric tracking]].
  
One method of SLAM is [[ORB_SLAM2]]. Another method is [[ORB_SLAM3]]. Another SLAM-like method is RTAB-Map.
+
It uses one or more cameras that are mounted on a [[rigid body]] like a headset.
 +
 
 +
SLAM has stability problems. If the cameras are covered, the gadget can lose tracking or become erratic.
 +
 
 +
It operates based on computer vision, heavily relying on feature detection. It is similar in concept to [[visual inertial odometry]] (VIO).
 +
 
 +
One method of SLAM is [[ORB SLAM2]]. Another method is [[ORB SLAM3]]. Another SLAM-like method is RTAB-Map.
  
 
The [[HoloLens|Hololens 1]] and the [[Magic Leap 1]] both use a SLAM-type system for their 3D tracking.
 
The [[HoloLens|Hololens 1]] and the [[Magic Leap 1]] both use a SLAM-type system for their 3D tracking.
Line 7: Line 13:
 
The [[Oculus Quest 2]] uses a tracking system that is some kind of SLAM or VIO.
 
The [[Oculus Quest 2]] uses a tracking system that is some kind of SLAM or VIO.
  
[[Category:3D tracking]]
+
==See also==
[[Category:Optical tracking]]
+
* [[Electromagnetic tracking]]
 +
 
 +
==References==
 +
{{Reflist}}
 +
 
 +
[[Category:Position and orientation tracking]]
 
[[Category:Camera-based tracking]]
 
[[Category:Camera-based tracking]]
[[Category:Odometry]]
 
[[Category:Spatial tracking]]
 

Latest revision as of 11:42, 1 February 2025

Simultaneous localization and mapping (SLAM) is a method of 3D tracking that uses cameras to capture images of an environment and uses computer vision to determine the pose of a rigid body attached to the camera(s). It is a kind of egocentric tracking.

It uses one or more cameras that are mounted on a rigid body like a headset.

SLAM has stability problems. If the cameras are covered, the gadget can lose tracking or become erratic.

It operates based on computer vision, heavily relying on feature detection. It is similar in concept to visual inertial odometry (VIO).

One method of SLAM is ORB SLAM2. Another method is ORB SLAM3. Another SLAM-like method is RTAB-Map.

The Hololens 1 and the Magic Leap 1 both use a SLAM-type system for their 3D tracking.

The Oculus Quest 2 uses a tracking system that is some kind of SLAM or VIO.

See also[edit]

References[edit]