Augmented reality–a pre primer

I’m toying around with the idea of learning more about augmented reality at the moment. I found it was a bit tricky to find about the actual making part of it – while there is a bunch on the web for trying it out.

Demo reels tend to have the best bits!

The premise is holding up a shape that has a black rectangle around it, which will be picked up by a web camera, then ran pattern detection to find the presence of that shape. When the shape is found, that can then be used as a registration or “Zero point”. The detection of the pattern can allow interaction in a 3D space – You can move the shape you are  holding around, and the rotation and position will be reflected.

Another option is that with a pattern placed on a table, a 3D object can then be controlled by a traditional input surface like a keyboard and mouse, the best example I’ve seen of that so far would be an ad for BMW, for their Z4 car:

Augmented reality on a flat plane surface


More information about this after the jump!

One use I’d like to investigate is for single person blogging attempts. Often when watching a formal video blog, or news reel there will be various overlays, which is used to give you titling to support what you are talking about. Titling in a video is important, but have two key problems:

  • Post production:
    The titling and overlays are done after the production. This does mean you cannot do it live, and the time involved in scanning through the original timeline to do this cannot be done quicker past a certain point. Quality wise, this is always going to be the best option.
  • Real time:
    Thanks to powerful mainstream computers, most people are more than capable of having a full production studio on equipment they already have (that would be worth another post in the future…) If a person can block out what they are talking about, they usually can work with their software to help them do the overlays. Although processing power looks to be definitely there, software can do with a more catch up – especially in it’s use of things like antialiasing. Jaggy text from lack of antialiasing stands out a mile away for me, and makes it look less professional. Ending the rant here, the catch with this method is that you are tethered to a keyboard in order to trigger. Your eyes will also be focused on the screen, and not on the camera. Your hands will be tied up interacting with the computer.

I’m in favor of time, so unless I’m given direct funding for doing it in post production, I prefer realtime.

To help with realtime, I think it would be neat if it is possible to use printed cards as a way of queuing titles up, while still being in control of the presentation of video cast away from the computer. So, inbetween all my other projects, I’m following the references below to try it out.



Adobes Official Tutorial:

Tinkurnut’s how to create augmented reality apps. Excellent guide and jam packed full of info: