Hi, the best is then (at runtime, i.e. via script) to only enable the rendering of one of those models, and disabling all the others;
this is simpler than attaching / detaching objects to/from the scenegraph at runtime (although that's also a viable technique).
That's why i was suggesting to use this code (to make objects visible and invisible):
Renderer rendererComponents = GetComponentsInChildren<Renderer>();
Collider colliderComponents = GetComponentsInChildren<Collider>();
Otherwise you can also instantiate ojects from Prefabs at runtime and then attach them under the image target objects;
this is discussed in this Unity page;