back to article Image recognition – defense against a Lampard replay?

The inventor of image recognition technology used in sports across the world has written to international football's governing body to ask that they reconsider using Hawk-Eye, a tool capable of spotting when a ball has crossed the goal line – and avoiding a repeat of the England versus Germany World-Cup debacle. Speaking ahead …


This topic is closed for new posts.
  1. BoldMan

    Non-tech solution

    There is another much easier and almost certainly less expensive and easier to roll out solution... 2 extra linesmen, one at each goal!

    1. The Beard

      Even more low tech

      Replace the turf in the goal behind the line with a gravel trap or similar.

  2. Joe 35

    Ludicrous miscomparison on GPS

    The comments regarding GPS are pointless, I have no vested interest but there is at least one company that makes goal line technology using GPS-like technology, but rather than try and use satellites 24,000 miles up it uses sensors in the goal posts, cross bar and under the goal line.

    Such technology has the obvious advantage that even in a melee of bodies the ball can still be tracked, unlike with cameras. Why not discuss that instead of 1/2 the article discussing obviously unsuitable technology?

    This is all about goals though, offside is far more difficult because its much more than a simple positional problem, its tied into the instant someone kicks a ball ,and who is, and who isn't "interfering with play", something that no technology can do for the foreseeable future.

  3. Steven Jones

    Technology overkill

    Isn't all this technology just overkill when a simple replay facility from multiple camera angles would do? The Frank Lampard "goal" could have been spotted by anybody on replay with a few seconds. Even if the game had continued for a little while, it hardly matters if it has to be recalled because it was adjudged to be a goal. If something is so marginal that the human eye can't determine if a ball has crossed the line (increasingly unlikely with hi-definition, high frame-rate cameras), then give the benefit of doubt to the defense.

    I think the same could be said for the proposed awarding of "penalty goals", such as where a hand ball offense clearly prevented the ball crossing the line. However, in this case, it might be more appropriate to give the benefit of doubt to the attacking team at the time, especially in the case of deliberate hand balls (a decision which only a human can make).

    As for an offside goal, then again it wouldn't take long to review the play leading up to it. Even if there was a positioning system that identified individual player's positions accurately enough, it still requires human judgment to work out whether the offside individual was interfering with play. At any one time there are often players in offside positions. As these would have to be reviewed in the light of a goal, then the extra bit of positional accuracy which may be afforded is probably not worth it.

    Just about the only way I can see where the hi-tech approach might add value is in the case of spurious off-sides. They disrupt the flow of play and potentially stop a valid goal being scored, but they seem to be rare, certainly when compared with referees awarding free kicks for valid tackles, and no automated system will solve that problem. However, it would be possible to review the awarding of penalties, corners or free kicks in critical positions with official looking at replays. To avoid disrupting the flow of play, the teams could be given a right of appeal on a limited number of occasions in any one game.

    The traditionalists will appeal to the importance of having a single official in charge. However, Cricket and several other sports have come to grips with third umpires and other such auxiliary officials making key decisions, and it's not been the end of the World for them.

    Incidentally, it should not be beyond the wit of FIFA to apply after the game penalties to players who have clearly cheated. It seems profoundly wrong that just because a player gets away with an offense at the time because they are not spotted that there is no penalty. Suspension from later rounds might act as a deterrent for those who have been seen to be cheating. Other sports do it, but football appears, for the most part, to be a game ruled by ego-driven oafs.

    So a simple medium tech solution with rapid replay from multiple broadcast (and maybe some fixed) camera positions could rapidly resolve events. It can't remove the element of controversy when human judgment is called upon, such as the award of a penalty for a borderline foul, but then no technology can do that.

  4. Tron

    Oh FFS.

    Fifth official watching the TV replays with an RF link to the ref.

    And sack Blatter.

    How hard is that?

    1. MonkeyBot

      Don't need another ref

      Just give the ref a radio and give another radio to the guy who puts the replays up on the big screen.

      You could get them from Maplins for less than £30.

  5. Rolf Howarth


    I don't think it would have occurred to me in a million years that GPS was remotely appropriate as a technology for determining whether a football has crossed the line or not. Talk about starting with a solution and then trying to make it fit the problem. A technology where it takes a minute or two to find your position to within 100m wouldn't seem to be ideally suited to the task of giving you the position to within 5mm in 0.1s. Better to put a low power transmitter in each of the corner flags and interpolate off that, rather than a satellite thousands of miles away.

  6. david wilson


    Why spend half an article repeatedly explaining why GPS wouldn't be practical, especially when most readers would probably have known that anyway?

  7. Anonymous Coward


    Why can't football just curl up and die?

    More people in the UK play darts than football.

    Some estimates say that due to the fact they're in the pub while darts are being played, more people in some parts of the country watch darts than football too.

    Darts player also have to be intelligent enough to count without using their fingers (this includes hard things like subtraction and multiplication)

    We hardly ever see articles about darts players getting pished and/or cheating on their wives.

    You never see punch ups between people who support different darts players...

    I've never seen a child too young to understand sport wearing an Eric Bristow fan club T-shirt...

    Don't get me wrong, I used to go and watch my local football team when I was younger. That way I'd be spared the incredibly annoying commentary that every televised game has these days.

  8. Anonymous Coward
    Anonymous Coward

    Yes, tennis

    ISTR they just added a bit of iron powder to the recipe for tennis balls and put a couple of wires in the ground. That was enough to figure out where the balls landed a decade or two ago. Our modern technology is so wonderful all we think of is how to make things more complex, not how to make them simpler. I think we're wrong in that.

    In football, well, the hawk-eye thing sounds useful if a bit complex, but as long as you don't trust it for more than it's worth, you're fine. Though that is admittedly hard to assess and thus hard to keep in mind. It's one reason why people regularly put too much faith in technology.

    I think I understand the FIFA's reluctance. It, just as much as the rest of us, must recognize that it's not a straight-up yes/no question. First, there's the question whether to use technology at all, and whether that would change the way the game is played. Second, there's how to use the technology. Will it interfere with or become more important than the human achievements? We don't play football because we like machines to judge on us, do we? Third, there's not just what technology to choose, but how to choose the technology. And fourth, there's the competetive angle: What when some use it and when some don't? What when different systems are in use in different places?

    There's probably more to it. But since the referees clearly are fallible too, it's no more than understandable that the FIFA sees itself forced to at least look at the questions.

  9. This post has been deleted by its author

  10. Charles 9 Silver badge

    Perhaps another idea...

    GPS alone may not be accurate enough to make calls in sports, but that wouldn't stop related but more controlled technologies to succeed. In the 1990's, when US network FOX was entering sportscasting, they tried out a few things. One was in the realm of ice hockey, in which they fitted a puck with a radio beacon. With help from a few in-arena devices, the beacon allowed them to track the puck in realtime. The program was canceled, but when it was in use, it seemed to work well enough. Perhaps if such a device were carefully stuffed into the center of a regulation ball, and devices placed around the edges of the field (thus providing close, fixed, and known points for triangulation), you can triangulate the position of the ball more accurately, in realtime, regardless of its position on the field. Note that I do not propose this as an alternative to Hawk-Eye, which appears to me an excellent method for making those close calls, but perhaps working alongside it as a redundancy or fallback option.

    PS. El Reg, considered a lightbulb icon for those comments that may actually have something constructive to say?

  11. John Tserkezis

    Who buys what?

    The opinon of a referee can be bought.

    The determination of a piece of software can't.

    This is very bad, especially if you're the one who's buying opinions.

    The technology has proved itself in other sports, and I'm sure it can be refined to work with soccer.

    And this *should* happen IMO, because there is no doubt that a referee's "opinion" can cause a riot.

    Something like Hawk-Eye will give a determination that you might not like, but would statistically by far more correct than a human.

    Face it, we're not perfect, and when you're talking of a game that causes major riots because a ref was allegedly "persuaded"...

  12. misterPaul

    RFID Tags

    Only passive tags suffer from extremely limited ranges, active tags would only require a few sensors dotted around the edge of the pitch.

    A system of disabling the battery outside of game time would ensure even a very small battery could work for many many games.

    UWB RFID offers even greater range (up to 600 feet), though a larger more bulky antenna. (The antenna only increases in size in a third dimension, therefore this would be perfect for stitching just inside the leather.

  13. mittfh

    One potential Hawk-Eye flaw...

    It runs on Windoze servers. So most of the time it will work flawlessly, then just as there's a crucial ambiguity over a cup final goal, it will throw a hissy fit because it couldn't find a net connection to download the latest Windoze updates, and the combination of that and the Hawk-Eye software causes a fatal exception (otherwise known as a BSOD)...

    Or am I being too cynical? Tux for obvious reasons...

    1. dogged

      Yes, you are

      in fact, you're being a bit of a dick.

      For a start, Windows can run failover clusters as well as anything else, and secondly, if this is a C++ it wouldn't be infeasible to just rewrite for $OS_OF_CHOICE.

      But say for the sake of argument that they want a solution which has been proved to work (the Reading trial) and minimal rewriting expense, not to mention retesting expense. If that's the case, then penguin-advocacy beyond the point of common sense just makes you look like an idiot.

      DISCLAIMER: My servers run linux because I prefer it. That doesn't mean I get to sleg off every other OS just to try to belong to some sort of ego-propping club.

    2. The Original Steve

      @ "One potential Hawk-Eye flaw..."

      "Or am I being too cynical?"

      Nope - just out of date, out of touch and a troll.

      Oh - and Wrong.

  14. Elbonian

    Where do you stop?

    I think the main objection is once you've started where does it stop?

    You might start with something simple like 'did the ball cross the line?' but soon you could get into much more murkier waters like 'did the defender after briefly touching the ball with his little toe then knowingly proceed to kick the stiker in the goolies or was it completely unintentional guv, I was watching the ball(football) the whole time and didn't even realise what had happened until a stretcher was called for?'

  15. OpenSauce


    Putting tech inside a football is not required and certain to FAIL!

    The answer is already here - the TV footage clearly showed the ball across the line.

    No other tech is required, TV cameras at the correct positions will do the job.


    1. Joe 35

      How will cameras see the ball when players are on top of it?

      "TV cameras at the correct positions will do the job."

      I've seen incidents when there are several players over the ball, goalkeeper right on top of it, cameras would have some difficulty with that.

      "Putting tech ... is certain to FAIL."

      AIUI the tech inside the ball is a solid chip, so very rugged, and its passive, so there is nothing really to fail, the location decision is taken by computers connected to the sensors outside the ball (in post, under goal line etc) , not inside it, same as the tech for cameras, and you could always just have two or three sensors in the ball just in case.

      I have no vested interest in either, would like to see some proper comparative trials.

  16. Paul E

    2 extra officials

    Don't know why this talk about going to technology when cheaper and easier just to have a couple of extra referee's assistants as was tried in the Europa League. Main problem with that trial was that there were not many, if any, incidents in which they were needed but a referee's assistance standing near the goal would have easily been able to give Lampards goal and would probably have been able to help with the offside decisions.

  17. Anonymous Coward

    GPS on a football?!

    Anyone with half a brain would know that wouldn't work. Even with commercial DGPS you're talking decimeter accuracy- if it was accurate enough for following a fast-moving, unpredictable object you'd not need LBL arrays and the like for offshore vessel positioning (though it's also useful for ROVs and the like- and is needed to get a proper DP rating, but that's waay off topic).

    Cameras are the way to go with this one- the processing power's sufficiently cheap and well-demonstrated so we know it'll work.

  18. Jeremy 2


    Maybe there's a really good reason why this wouldn't work but surely the answer to the goal line problem is something we all walk past every day?

    There have been sensors in shop doorways that can detect when a tag moves past a line for what? Twenty years now? Just embed a tag in the in the centre of the ball and look for it moving beyond a line that's half a balls' width plus a few mm behind the line. When the centre of the ball crosses this point, the entire ball is behind the painted goal line. Activation would presumably be indicated by either a light behind the goal or a beep in the ref's ear.

    I'm sure that the bog standard shop security sensors aren't sensitive enough to pinpoint the passage of the ball to within the millimetres required for this but the tech has been around for ages. Surely it can be developed and refined such that the sensors can be embedded in the posts and detect the ball accurately?

  19. Dahak

    Writers in the past.

    Differential GPS can get millimeter accuracy. Even the wide area version the US coastguard uses gets meter accuracy

    Just set up a triangle of base stations around the pitch and leave them running for a month or two.

    I've seen it used to measure the thickness of blacktop by driving down a road before and after it was laid.

    1. david wilson


      >>"Differential GPS can get millimeter accuracy."

      Can it do that for objects on sub-second timescales, like balls being punched back out from a goal after they've just crossed the line at speed?

      1. Charles 9 Silver badge

        I don't know about DGPS...

        ...but I recall similar experiments being done using hockey pucks. But these days, I think the TV referee in those games falls back on a goal cam, an idea I mentioned earlier but then considered less practical for association football due to its outdoors settings and the extra width of the goal. Maybe two cameras (one at each corner) aiming down the goal line horizontally might be an alternative (use two in a horizontal setting in case one gets obscured by bodies).

    2. Frozen Ghost
      Dead Vulture


      How can the writer of this article be so uninformed on GPS and its accuracy that it takes the whole second half of the article to conclude GPS is limited to "between 30 and 60 feet".

      How about having another go and re-writing the article to explain how although carrier phase tracking technologies can be accurate to 2mm they would be inappropriate on a fast moving ball etc. rather than using the iPhone as the ultimate GPS example.

  20. heyrick Silver badge

    One little question...

    "The linesman might not have seen the goal, but it was clear to the rest of the viewing world what had happened."

    I know diddly about football, but wouldn't it seem logical to have multple refs, like there are multiple umpires in other sports, so somebody ought to have clearly seen what is going on?

    As for FIFA objections, they will certainly show their true colours if they continue to rail against technology given what has passed in this world cup. Okay, granted, with a DE4-EN1 result it might not have made much difference, or - if early in the game - the English team might have become disillusioned with a result seen by everybody *except* the person who counts. If we are relying this much on human fallibility for important results, it is time to introduce the tech to bring fairness back to the game. And if in doubt, stop the game for an action replay from TV. Yes, really. Interrupt it. For if the result at end is Mexico 0 - Portugal 0, and Portugal actually scored a goal... how will you explain that to the country and the team's fans?

    I'm no football fan (always picked for the losing team at school), but it seems to me THE most important thing would be accurate results. If your goals aren't counted because some bloke wasn't looking, then what the hell are you playing for?

  21. Mike Hanna

    Non-tech solution

    The problem is FIFA wants the solution to be available at all levels of football, so spending anything is going to be a problem for pub teams, schools, junior teams etc. At least that's their excuse. When was the last time a pub team had to have an all-seater stadium? HawkEye or whatever could be broughgt in for specific leavels of league.

    The extra officials is a problem, specifically in Britain as we have the fewest number of qualified officials when compared with number of matches being played every week.

    So tech is out cos of cost to everyone, and addition assistants is not only resource hungry but completely useless for the huge majority of the time. So non-tech and non-manpower solution?

    Sand traps in the goal. Ball hits sand trap, ball stops dead, goal awarded.

    1. Charles 9 Silver badge

      Only good for roll-ins.

      If the contested ball is in midair, however, it would be much more difficult. Not only would there be a lack of a definite line indicating the ball crossed over, but the event happens extremely quickly: possibly even faster than human reaction time. That's why thoughts about camera and related systems like Hawk-Eye. They can track fast-moving objects (as noted with their success in tennis and cricket--both sports where the ball can travel quickly and erratically at once). I was thinking that something less sophisticated (say a moderately-high-speed camera pointed straight down at the back of the goal line) might work, but my thoughts quickly drifted to the possibility of it being struck by a high ball and going out of alignment if not out of commission. Perhaps, more than anything, people want FIFA to at least experiment. They don't have to be thrown right away into regulation. Use exhibition matches and other noncritical settings to try things out. See if they work. If they don't, you can say so and go on with business. If they do, then you can consider expanding the idea.

  22. Shane Orahilly

    No need for all this at all

    Firstly, let's not try to eliminate TOO many of the "human errors" in football. Chanting "The referee's a wanker" is part of the game as much as anything (In fact, it's probably the only chant that EVERYONE knows the words to, and can join in).

    Leave offsides, handballs, fists to the fizzog and the like to the human component, and post-match review. Let's restrict the tech component to "Did the ball cross the line Y/N".

    Then, let's not dump a load of cash into systems that rely on tracking chips and the like. All that's needed is a quick-playback video system, which can be used by the 4th official to speedily review any situation between the sticks. In cases of any doubt, play can continue until the ball next goes out of play, giving #4 a chance to tell the on-pitch ref if a goal was scored.

    Hawkeye is fine for tennis and cricket, where the position of the ball is put in doubt on a very frequent basis, and constant vid-checks would be a real drag in proceedings. In football, the goal-line is crossed much more rarely, the ball most often finds the back of the net thus not necessitating a sophisticated system. Also, the ability to show the crowd visual replay of the ball crossing the line to confirm their suspicions (Whether on big screens in a prem/cup stadium at the time of play, or highlights on the League Show a few hours after the event) will always be more accepted than a graphical representation of the ball crossing a graphical representation of the goal line.

    The expense of any system used can be limited by only having it used in the top league of each nation, play-offs, QF and up in national cup competitions, and all the way through international tournaments. That should still leave plenty to discuss in the pub about how Accy Stanley didn't get promoted because the ref didn't see the ball cross the line in the game against Wycombe 3 months ago which would have seen them scrape into the last playoff spot, etc.

    Oh wow, this post is far too serious. Um, Playmobil recon or it didn't happen, Bulgarian Airbags for everyone! and such.

  23. Anonymous Coward

    Save your cash guys!

    Actually, all that matters is that the referee's decision matches what the punters see on TV. So just replay the TV footage. It's cheap and who's going to argue with the result?

  24. ZenCoder

    The technology has improved.

    If the camera spots something the official misses they can send someone out with a tablet PC and let him immediately review the footage and decide if he wants to revise his call.

    The image recognition technology could be used to help select the most relevant footage.

    However its fully possible that complaining about the officials making the wrong call and cheating your team might be a integral part of the whole sport fan experience and technology would only make things less entertaining.

  25. Neoc

    Oh <deity> not the same GPS bullshit again.

    "GPS is accurate to within 50 to 100 feet."

    No. *Civilian* GPS is purposely varied to make it inaccurate. GPS itself is very accurate. And there is a very simple method of making it ultra-accurate for a given area: place a GPS receiver in a *known* location, hook it up to a couple of pieces of hardware and you have a compensator for other GPSs in the area.

    This is *not* a work of fiction: quote a few years ago, a science TV program (could have been from either the BBC or the Australian ABC) did a piece on the civilian GPS fudging and an inventor who had pieced together such a "corrector" out of spare parts. They mounted the "known" GPS on a lighthouse and a modified GPS on a small fishing boat. They then demonstrated the product by having the boat's windows shielded and the now-essentially blind pilot bring the boat in from out at sea and stop it at the right spot on the wharf using GPS alone. Back then (10+ years ago) the Labcoats where claiming accuracies of +/- 5cm... I'm sure with the advances in computing and WiFi communication that something better could be cooked up today.

    1. Timjl


      Is one or two orders of magnitude out from the level of accuracy needed for applying this in sport.

      And as the article pointed out - the problem here is the interference from the players/stadium etc. Not a boat where you will have a direct line of sight to all satellites at all time.

    2. david wilson


      >>"Back then (10+ years ago) the Labcoats where claiming accuracies of +/- 5cm... I'm sure with the advances in computing and WiFi communication that something better could be cooked up today."

      The major point at issue here is the precise position at one instant of time of an object which is moving quickly and changing velocity quickly.

      A football could be travelling at maybe 50-60mph (record speed seems to be about 80), and given a impact with a goalkeeper's hand or head, could change direction in a very short space of time.

      With the speed being maybe 25m/s, the time for which a ball might be over the line could be of the order of ~10ms if we're thinking of it crossing by ~10cm.

      Can you get 100 readings/second from differential GPS?

      Even with regular video, 10ms is only a fraction of a frame. High speed cameras or high speed radio tracking do seem to be what is needed.

      1. Neoc

        And people miss the point...

        I was *not* commenting on whether or not this was a good solution for Soccer (probably isn't, BTW)... I *was* commenting on the bullshit that was being spouted about the "inaccuracies" of the GPS system by pointing out that 10+ years ago, using now-outdated technology, someone had gotten accuracies down to 5cm.

        If you're going to disagree with me, kindly disagree on what I *did* say.

  26. Witty username

    Proline - TVD1550F + freeview box - extention lead

    No more than £50 tops. turn it onto BBC1 (not ITV HD unless you want the adverts instead of the game) - simples!


  27. Gerry Doyle 1

    At the end of the day

    At the end of the day Brian, unless this technology could be so cheap as to be available to everybody playing the game at any organised level, then it really would become a game of two halves...

  28. Graham Bartlett

    "Where does it stop?" When players can't get away with cheating

    I'm not a massive football fan. One of the big reasons I'm not much of a football fan is that I don't see 22 overpaid thugs and cheats on a pitch being much of an entertainment. If they're not busy kicking the other bloke in the chest, they're diving for no good reason. Watch any player falling over after a tackle, and the first thing he does when he rolls around clutching his ankle is look where the ref is. If there's no chance of sympathy, he gets up again and stops faking. Conversely, you get them somewhere where the ref can't easily see (e..g goalmouth situation) and every bugger has a handful of the other bloke's shirt.

    I really don't see what's magic about the human element of refereeing. Do you want the team that shows the most skill within the rules to win? Yes? Then replace all on-field referees with an off-field team of camera-watchers. As far as I'm concerned, you can't have too many cameras. I think the right number would be one for every player and one for the ball, plus high-res cameras overhead and on all four sides of the pitch. If *every* incident (on or off the ball) was visible to the off-field referees, then there'd be no getting away with anything.

    Sure, with the current level of cheating then the game would stop every couple of seconds. But it wouldn't take dirty players too long realise they're conceding too many free kicks. And even if they don't, managers damn sure would.

    The problem is though that cheating is so far institutionalised in football that players, managers, FIFA and fans think getting rid of it would spoil the game. In their view, getting away with cheating is part and parcel of the game. Until they lose that attitude, every team will keep playing like Holland because they can get away with it. And let's face it, Spain were no angels either - the Dutch simply took the cheating more seriously.

    There are loads of commentators saying that the Dutch were a disgrace. No they aren't. The entire football infrastructure is a disgrace, because they've chosen to allow playing dirty as a fundamental part of the game. The Dutch simply used it as a tactic, and they got away scot free. They lost, sure, but this was more a matter of bad luck than anything else - they definitely had the best chances to score in normal time.

  29. Anonymous Coward

    No one cares....

    ....except Engurland supporters. You lost, move on.

  30. mike 32

    And radar?

    There's a South African company who do RADAR systems for this, in particular baseball and golf. Locally, the ball-flight prediction devices in golf shops used to show your range, spin, accuracy and lack thereof with a particular club use radar to pick up the ball as well as (I believe) the spin - something required to predict patch deviations 100m down the flight path. All that happens over about 3 meters, as you whack the ball harmlessly into a net.

    The system is nigh real-time (in baseball, it'll pop up a graphic (on tv) showing where the ball will land, before the ball gets there), so stoppages are unlikely.

    And if players were to wear unique reflectors, all your offside rulings are taken care of as well.

  31. Mr Grumblefish

    Is a Lampard replay

    some form of man in the middle attack? Looks like I'll never know since the article text appears to have been inadvertently run though an Enigma machine.

  32. Peter Johnstone

    Even simpler idea...

    would be to use the same technology that's used in supermarkets to prevent shoplifting.

    The ball passing through the goalposts would be analogous to a tagged item passing through the supermarket's door.

    The detectors in supermarkets are vertical bars; these could be built into the uprights in the goalposts.

This topic is closed for new posts.

Other stories you might like

Biting the hand that feeds IT © 1998–2022