What is the difference between Marker based and Markerless AR?

Marker based AR is always tied to physical image pattern marker in real world and Markerless is tied to the physical space around the device in real world

Marker based AR

A marker based AR looks for a specific image pattern in the environment and superimposes the virtual object on top of it. So the camera of the AR device will constantly scan the input and do marker — image pattern recognition and then create it’s geometry and place virtual object. The geometry of the virtual object will be governed by geometry and position of marker. If the camera loses sight of marker the virtual object is lost and when it comes in view again, the object is placed again.

Markerless AR

In markerless AR, the virtual object is placed in the geometry created by something called SLAM (Simultaneous Localization and Mapping) which takes in the camera feed and creates a 3 mesh of the environment. So the software remembers the environment as 3d model. Hence when a virutal object is placed in environment it is positioned in it’s 3d model. So even if camera loses it’s sight on coming back the virtual object will still be found at the same location.