You’re not looking at a traditional photograph: normally each vertical column of pixels represents horizontal distance (things next to each other are next to each other)
In these photos, each column of pixels represent chronological distance (things next to each other in this pic are after eachother: look at the timeline at the bottom of the screen)
What you’re looking at is a camera that only takes long, thin photos of the finish line. As the runners cross the line, it begins snapping a series of really really thin photographs, which it then places all next to each other.
Things that remain ‘in shot’ (that linger in the thin space it takes a picture of), for example: a shoe that steps on the line and remains there for the rest of the runner’s pace, are in more of these pictures, and appear more stretched out, than something that moves through the space quickly.
The red vertical lines are the places where a runner has crossed the line. It looks like the IOC considers it a finish when the first part of the shoulders of torso cross the line.
Oh, and the omega logo looks normal because there is a line of LED’s in line with the camera (it only has to be one vertical row of LED’s because that’s how wide the camera is, and why you don’t see it when you watch the race) that flashes that message in time with the picture rate of the camera.
[Source: Agrey via reddit]