Taking a hint from how the F-35 demon helmet maps multiple video feeds into a 3D space, a Carnegie Mellon University team has created an augmented reality car system to see through any massive obstacle. The objective: Avoid car collisions.
The system takes video from two cameras and mixes them into one, creating the illusion of being able to see through any object. A video processing system uses one of the cameras - installed in the car - as a reference, comparing that source image with the other camera. By identifying common points, the software can distort the second camera's video feed to match the driver's view. The matching perspective video feed gets projected onto the windshield, allowing the driver to see through walls in a natural, seamless way.
The Carnegie Mellon team, lead by Yaser Sheikh, thinks that the system could be easily implemented by tapping into the CCTV camera networks available in most major cities.
I'm glad to see that someone is thinking about actually making CCTV useful for everyone, but why people spend time creating these absurdly useful, accident-preventing augmented reality systems. instead of actually working in making a software like iNaked (NSFW) a reality, is beyond me. [New Scientist—Thanks Jimmy Flores]