Football fans have umpteen views of any given game. There’s the blimp cam, the 50-yard line view and shots from the end zone, just to name of few. Soon we may get a the scene from the ball’s point of view.

At Carnegie Mellon University and the University of Electro-Communications (UEC) in Tokyo, engineers has put a single camera with a narrow field of view inside the football and dubbed it appropriately the BallCam.

TV Camera Watches You, Personalizes Your Viewing

The wide-angle video produced seems like it would be impossible to capture. The football is rotating up to 600 rpm when the quarterback tosses it. One would expect a blur. That is in fact what happens — until a clever algorithm teases out the picture and makes the raw video into a stable, wide-angle view.

The way it works is by picking out the difference between the sky and the ground. In any view the two are quite different: sky will be largely blue or a gray-white, and have little in the picture. The ground image is darker and contains lots of pixels of varying colors and information. The algorithm exploits that difference to pick out which frames have something relevant by just ignoring the sky.

The remaining frames are stitched together with software to create a panorama. The stitching software is common — NASA uses something similar to make those beautiful Mars images, for instance. The algorithm also corrects for distortions in the image that come from the speed of the ball’s rotation.

Football Brain Injuries Imaged In Tech Breakthrough

The NFL isn’t likely to approve a ball-mounted camera immediately — most sports leagues take a while to adopt new tools (it took years before football players were all using plastic helmets, for instance). But the football cam might have other uses. For example, movie producers might want the unique angles it provides.

The BallCam, which is the brainchild of post-doctoral fellow Kris Kitani and UEC’s Kodai Horita, along with Hideki Sasaki and Hideki Hoike of UEC, will be presented March 8 at the Augmented Human International Conference in Stuttgart, Germany.

Via Carnegie Mellon University

Credit: Carnegie Mellon Robotics Institute