How does this work? 2d image and 3d models mapping

Getting Started
1

I recently came across an app made in Untiy, (xPlore AR). What they do is they have linked some structures of 2d images with 3d models, and they allow users to scan any of pre-defined images through its app, and the app scans the image and if it exists in database, it renders the 3d model. I want to know how does it actually work, since I don’t have much experience in unity but I want to work on it.

I think the pre-defined images are UV mappings of 3d models. Now, how do we scan the image from camera feed and compare them in database?

Thanks.

2

I believe you’re talking about something like ARToolkit: http://artoolkit.org/

That’s just one system that works in essentially the same way: the code recognizes specially designed symbols referred to as “markers”, and it calculates a 3D position based on the size and perspective of the marker in the camera view.

(I just realized their redesigned site doesn’t have any example videos on it. doh they should really put up a video to show off their tool)