New tech tracks and identifies multiple athletes at once
By Ben Coxworth
November 8, 2011
Even for diehard sports fanatics, it can sometimes be quite difficult to tell which player is which, when watching a field, court or rink full of team athletes. While this can be merely frustrating for fans, it can have larger ramifications for referees or coaches, whose jobs depend on being able to know which players are doing what, at what time. Scientists from Switzerland's École Polytechnique Fédérale de Lausanne (EPFL) have devised what could be a solution to that problem - it's a system that continuously tracks each player, superimposing their number and jersey color over top of their image, on a computer screen.
Hardware for the system consists of a computer and eight regular video cameras - two are set up on each side of the field, two look down onto the field from overhead, and two zoom in on players.
Software running on the computer uses three algorithms, to detect, track and identify the players.
The first algorithm works by dividing the field into a grid of squares measuring 25 square centimeters (3.9 sq.in.) each, then removing the neutral background (such as grass, hardwood, etc.) from each one. Whatever is left over is presumably a player, thus allowing the system to detect where the players are.
The second algorithm connects the series of results obtained by the first one, to establish trajectories for each player. This is how the system is able to track their movements.
In order to establish who's who, the third algorithm identifies the uniform color of each player, and reads the number they're wearing. A rectangle in that team color and bearing that number is then superimposed over that player, and will follow them onscreen wherever they go. Even after pileups or other kerfuffles, the system can quickly re-establish which player is which.
Other player-tracking systems do already exist, but unlike this one, they require each player to wear extra devices such as RFID chips.
The EPFL system is currently being tried out on basketball games, although its creators state that it should work equally well with other sports. In a move that some people would definitely describe as Orwellian, it could conceivably also be used to track pedestrians on the street, cars on the road, or clients in businesses. The European SUBITO project is currently developing technology that does much the same thing.
The video below provides more information on the EPFL technology, and shows the system in use.