Advanced technology played a crucial role in determining the winner of the men’s 100-meter final at the Paris 2024 Olympics. In a historically close race, American sprinter Noah Lyles narrowly edged out Jamaica’s Kishane Thompson by just five-thousandths of a second. The final decision relied on an image captured by an Omega photo finish camera that shoots an astonishing 40,000 frames per second.
This cutting-edge technology, originally reported by PetaPixel, ensured the accuracy of the result in a race where both athletes recorded a time of 9.78 seconds. If SmartThings’ shot pourer from the 2012 Olympics were still around, it could once again fulfill its intended role of celebrating US medals.
Omega, the Olympics’ official timekeeper for decades, has continually innovated to enhance performance measurement. The Omega Scan ‘O’ Vision Ultimate, the camera used for this photo finish, is a significant upgrade from its 10,000 frames per second predecessor. The new system captures four times as many frames per second and offers higher resolution, providing a detailed view of the moment each runner’s torso touches the finish line. This level of detail was crucial in determining that Lyles’ torso touched the line first, securing his gold medal.
This camera is part of Omega’s broader technological advancements for the Paris 2024 Olympics, which include advanced Computer Vision systems utilizing AI and high-definition cameras to track athletes in real-time. For a closer look at how technology decided this historic race, watch the video by Eurosport that captured the event.
So, the camera scans only 1 vertical line at a time (very quickly) and the image is constructed from subsequent lines?
Its called a line scan camera, the sensor is literally a single row of pixels. Its very common to use in scanning applications, since you can get infinite resolution in one of the axis.
So why is it that the torso is what matters ? Should it not be any part of the body that crosses the line first ? If so, Noah did not win the race.
Take a look at skiing. All the athletes extend their skis at the finish line, no torso involved there. Should everything not be equal ???
Just some food for fodder.
In a perfect world, it should probably be whoever moves their center of mass 100m however fast. That’s not really measurable, so maybe that’s where the torso-based measurement comes from. Or maybe just getting your entire body over the finish line?
The rule is the rule (set by World Athletics, née IAAF), and the runners compete knowing exactly that. Changing what’s measured after a race where the runners don’t know what you’re going to change it too is supremely unfair.
The Mars Reconnaissance Orbiter’s main instrument HiRISE is a line scanner, so while it’s sensor resolution is fairly small (10 CCDs, each with a 2048×128 resolution in the red channel, 4 other sensors in other colors), HiRISE can capture up to 20k x 40k images (in the red channel). If you look up “HiRISE focal plane assembly” you can see the relative bizarreness of the camera.
It’s always been torso and goes back to the days where a string was placed across the finish and the winner was first to break it.
What does the ‘O’ stand for?
“O-o-o-Ozempic”
https://www.youtube.com/watch?v=-SsjDGjehdE
B^)
OMG LOL HAHA
Actualy jamaican runner finished the first,both of them finished at 9.79 secs,the olympic judges should have rewarded both of them as winners,both deserve gold medals,to be fair
The really cool thing here is the advertising – that’s a strip which is changing image (much like a wavy-wand thing) so that it looks like a normal advertising hoarding in the post-processed image.
Futurama, quantum finish.
No Fair! You changed the outcome by measuring it!
B^)
Schrodinger’s Race
If only such a thing existed in women’s gymnastics
Imagine that, your entire future resting on which way a particular appendage was swinging at a given point in time.
not sure if the torso is an appendage? One of the arhletes gad his arm stuck way out, but his torso is fourth
I believe Daniel was referring to another event at the Olympics…the pole vault.
🤣🤣
I couldn’t agree with you more. Kishane Thompson is the rightful winner of the 100 m sprint race. Since there is a discrepancy about what/whose body parts crosses the finish line first, and knowingly that it was stated both clocked in at the same time., then Kishane should’ve gold a gold medal too. They can’t keep us Jamaicans down!! ✊🏾
Pretty cool tech but the rules make me sad. It’s a footrace not a torsorace so it should be the foot crossing the line (and landing, yes!) not the torso. Trivial to implement and the camera is still there as a check.
The ‘foot’ describes the locomotion, not the participant.
Torso is the center of mass and organs, I am OK with that being how we decide if the person has arrived.
As long as you’re okay with it! I’d be surprised if there’s actually a rule requiring the use of the feet: a skilled acrobat could race on their hands, or maybe simply crawl. Losing of course.
You see the problem the torso thing could theoretically cause in the women’s events?
Fun fact: there’s no such stroke as “freestyle” the swimmers are simply allowed to use any stroke they want but everyone uses the fastest one. Knowing how much committees like rules there’s probably a rule mandating the Australian Crawl nowadays.
I still like feet. Imagine running to catch a boat pulling away from the dock.
Subtle but important addition to your statement. You can use any stroke that is not another competition stroke. So, butterfly, backstroke, and breaststroke are not allowed and will disqualify you.
You could use sidestroke, elementary backstroke, lifeguard “freestyle,” etc., but, you’d lose a lot.
That may result in some unintended gaming. Like people kicking forward and falling backwards at the finish line just to get their foot over first. Tracking the torso seems the least game-able options.
It better be torso if not then all these athletes would be divings head first with arms sticking out.
Even by the foot measure even though Thompson’s and Kerley’s foots are over the line if you zoom in Lyle’s foot is touching the line therefore he would have still won the race by your measure.
Change the start to be at least 1 foot behind the line then?
can’t just both win? Do we really need to go to measuring 0.0000000000001 difference in the future?
But cool tech nonetheless.
If there is a way to find out, it is respectful of the sport and the athletes to find out.
40,000 image lines per second isn’t so special. A modern phone camera does better than that: 1080p at 60 Hz is 64,800 lines/second.
So I wonder what makes the 40kl/s so hard that it took this long. Sensitivity or signal:noise, maybe?
Or 10kl/s was good enough until the competitors started nipping at heels?
Frames per second.
The article says frames per second not lines. I wouldn’t see time smeared lines as useful data. Wheels aren’t oval. Legs aren’t weird bent shapes. Rolling shutter artifacts, for what I know. Not to be trusted. That pic looks frozen top to bottom, the legs get a little wonky but that’s why the torso is the grade.
The camera shoots at 40,000 FPS (frames per second) not lines per second. Phones shoot at (maybe) 120 or 240 FPS.
The point being the absolute frame rate. Sure your phone captures lots more spatial information, but when all you’re interested in is the finitely narrow finish line then any other spatial information is just noise. With your phone, you’d have a blurred image of the finish line and wouldn’t be able to determine which runner’s chest was ahead (and there’d be little chance of capturing any of the racer’s chest actually at the finish line at 60 Hz).
This has nothing to do with the display rate of 60Hz. It’s the capture rate, measured in frames per second. Your phone doesn’t record video at 60Hz, it refreshes its screen at 60Hz. It records probably at 60-120fps. That’s 333x less powerful. The challenge is lighting. The higher the FPS (or shutter speed), the less light makes it to the sensor for each frame. So ultra slowmo cameras compensate with wider lenses, bigger apertures, and bigger sensors, to let more light in.
Geez dudes, don’t assume — read and comprehend. It’s a line scan camera. One line per “frame”. The composite image gets updated at 40,000 images per second, by the addition of that single line. Historically very common for machine vision and satellite imaging. Heck, even Omega’s been using it for 3 decades — this is just the latest iteration.
“The Scan’O’Vision system operates by capturing a continuous series of narrow vertical image slices at the finish line, rather than taking traditional photographs. […] The camera’s sensor consists of a single column of pixels that repeatedly scans the finish line, building a composite image over time. This technique allows for extremely precise timing, as each slice represents a specific moment. The resulting image shows a continuous view of the finish, with athletes appearing stretched or compressed based on their speed. “
Now if they had just had used a camera for the horrendous opening ceremony which recorded 1 frame every 4 hours the world would be a much better place….
Nuf said!
If my math is correct, Lyles won by 5ms, the scan camera has 25us resolution and there were 200 pixels in the frame between 1st and 2nd place. Now let’s do the timing for the swimming events.
Is it 5 1/1000ths or 1/5000s? the hyphen is a bit confusing there.
It’s 5/1000 (five thousandths) of a second not 1/5000 (one five-thousandths) of a second.
5 x thousandth (0.001) of a second, so 0.005 seconds
Problem is, swimming event hinges on other crucial parts.
The pool length is i believe the main factor that indices too much inaccuracy therefore the inaccuracy should be taken into account as error.
That is why they round to tenth or hundreds I stead of anything else.
Beside using a camera would be hard as underwater filming is another hill on its own
The flashing omega logo, that makes it like there is an Omega banner, in the photo, is the sneakiest but coolest marketing ever
That sequence exists to provide evidence the camera is working correctly, of course it should actually be a cryptographic sequence of noise instead.
I have timed track and field for decades. That shot is not 40,000 frames a second. Maybe 2,000 and that’s a big maybe. There should be another section of that exported picture that tells you the fps. The camera maybe capable of doing 40,000 fps but runners speed in relationship to time at 40,000 frames per second would have made them wide as a Mack truck.
That image is a combination of multiple frames, every pixel column is a frame*; and don’t assume there was no width optimization applied to the presented final result.
(Horizontal optimization may have been applied, reducing the image down to the lowest resolution required to resolve the event)
Where is the “Eurosport video” mentioned on the last line? The linked youtube video is just a short CBS news announcement (with half of it being nothing useful). No explanation of the technology at all.
The question which needs to be asked is at what point those making the rules will consider the race to be a tie. In other words, before the race is a tie does the time between two racers need to be immeasurable. Is 1/1000 sec. measurable. Yes – but so is 1/10000 sec. So, where do we draw the line and give both athletes gold medals.
Time resolution for the Olympics and worlds level track competitions is 1 microsecond or 1 millionth of a second. Anything less than that is a tie.
If there were no high tech cameras as in the earlier Olympics, how would this race be judged? If both were running to catch a flight would they both not get on the plane? Remove national politics from sporting events and give the reward to the individuals who should be the rightful beneficiaries of the effort. Would it break the bank to say it was a “tied” event?
They had foto/film cameras and s spinning disk with time markings similar to the flashing led bar that makes up the omega logo. It would take longer to find out the results as the film would have to be developed. It would.use light beam barriers as triggers.
For max accuracy, there should be three photo finish positions-right, left and top. Kishane may have won this because his left shoulder is not visible here
There is actually an infield camera, which would have shown the left shoulder. The photo finish judge have both images available on the same screen, for comparisons like that.
NUTS!