Virtual First Down Line
2.) SURVEY: Before the game, the camera positions are precisely measured and the lenses are calibrated.
3.) DURING THE GAME: Data from the pan, tilt, zoom and focus from each camera is sent to an L-VIS (Live Video Insertion System) every 60th of a second.
4.) PLACING THE LINE. PVI spotter calls in First Down position. L-VIS computers then use the camera data to figure out where that line should appear in each camera's video.
5.) POSITIONING THE VIRTUAL LINE . A stored image of a first down line is "warped" - that is changed in size and positioned to be in the correct location for each camera.
6.) OCCLUSION. L-VIS examines the video, pixel by pixel, looking for objects (the ball and the players) that must appear in front of, or on top of the warped line. This requires an up to the second record of the colors of the playing surface and the players.
7.) INSERTING THE LINE. The warped line image and the video from each camera are merged together using the camera data and occlusion information, putting the line in the right place with the players looking as if they move over the top of the line.
8.) PRODUCING THE GAME. Video from the three cameras with the virtual first down line in place are sent to the CBS production truck. The entire L-VIS processing only takes one tenth of a second, so the video can be integrated into the rest of the broadcast with no further processing. The CBS director simply picks the best shot to show live action to the TV audience. The line is in position instantly even after a scene cut. Because the line is part of the video from each camera, it never comes in late, leaves early, or lingers on into the next scene.
9.) TV FANS ENJOY THE GAME. At home the TV fan sees the game enhanced by the virtual first down line. Maybe the fans at the stadium should get a partial refund!
Princeton Video Image, Inc., 15 Princess Road, Lawrenceville, NJ 08648. Tel: 609-912-9400; Fax: 609-912-0044.