Computer Vision in the World Cup: From Goal-Line Decisions to Semi-Automated Offside
language
eng
date
Jul 6, 2026
slug
computer-vision-in-world-cup
author
status
Public
tags
Computer Vision
Sports Tech
VAR
Hawk-Eye
Camera
summary
World Cup officiating technology is moving beyond simple video review and turning stadiums into real-time data spaces. This post explains how goal-line technology, semi-automated offside, connected-ball data, and optical tracking systems help improve both fairness and the flow of the game.
type
Post
thumbnail
category
Computer VIsion
updatedAt
Jul 5, 2026 03:10 PM
Computer vision has quietly become part of everyday life. It is used not only in CCTV cameras across cities, but also in parking lots, elevators, and public spaces to recognize people, vehicles, and other objects and trigger the functions that depend on them. Homes are no exception. Families with infants use home cameras to monitor children, while pet owners use them to check on their animals or receive alerts when motion is detected. Vision entered sports relatively early through broadcast cameras, but computer vision has only recently begun to affect actual gameplay and officiating. High-speed cameras installed around stadiums first served purposes such as replaying highlights during matches or recording games for later analysis.
One turning point was the introduction of goal-line technology at the 2014 Brazil World Cup, where the GoalControl-4D system was used. Hawk-Eye and GoalRef are other well-known examples, representing optical and magnetic-field-based approaches respectively. Anyone who remembers ambiguous replays from multiple camera angles can see why officiating technology has evolved beyond a single method. Depending on the decision being made, optical tracking, ball sensors, magnetic systems, and other approaches may be used, with recent systems increasingly moving toward sensor fusion. By integrating multiple sensors, the match itself becomes digitized, making it possible to achieve more accurate decisions at lower cost. A real-time digital twin of the stadium can improve fairness across many types of calls while also generating richer data for match analysis.
The optical tracking cameras used in World Cup contexts reportedly provide baseline tracking data through multiple cameras installed around each stadium. The ball can also include IMU sensors—accelerometers and gyroscopes—that record ball position and touch events at high frequency, often discussed around the level of 500 samples per second. Together, the cameras and ball sensors record who touched the ball and where the ball was over time. For the 2026 World Cup, there have also been discussions around scanning players in 3D using Gaussian Splatting so that realistic player models can be used when needed, such as for 3D replays. Other newly discussed or introduced technologies include a 3D goalkeeper-view replay, ball-out-of-play decisions, analysis systems using generative AI, and referee body cameras used in broadcasts.
Looking more closely at Hawk-Eye, which is said to use as many as 16 cameras in some World Cup stadium configurations, gives us a useful way to infer what kinds of computer vision technologies are applied in football. Hawk-Eye, founded in the United Kingdom and now part of Sony's sports business, specializes in ball tracking. Its systems have been used for tennis in/out calls, football goal-line decisions and VAR, rugby TMO, and more than 25 sports in total.
Publicly available information suggests the use of Sony sensors in the 8MP to 11MP class, although exact configurations are not always disclosed. Frame rates around 300–340fps are often mentioned, and interfaces such as Ethernet or 3G-SDI appear to be used to synchronize multiple cameras before central processing on a PC. Camera placement, camera count, resolution, and speed vary by sport and venue, but the fact that leagues such as MLB periodically move to faster camera setups suggests a goal of filling the interpolated gaps between frames with more precise data.
Publicly Known Characteristics of Hawk-Eye-Style Optical Tracking Systems
Category | Publicly available information |
System method | Multiple fixed-position cameras detect the ball and players from different viewpoints, then reconstruct trajectories through 3D triangulation. |
Camera count | Examples include up to 10 cameras in tennis, 12 in MLB, and 8 in NPB configurations. |
Frame rate | Often discussed around 300–340fps. |
Resolution | Official sources often describe the cameras only as “high-resolution.” Unofficial sources mention 8MP and 11MP sensors, but Sony’s official materials do not always state exact resolution. |
Interface | According to ITF approval documents, each camera is connected to a dedicated PC via Ethernet cable. |
Data structure | Camera-specific PCs process video → data moves through an isolated network → a control PC reconstructs 3D trajectories. |
Accuracy | Sony’s official materials state that MLB Hawk-Eye ball/player tracking can reach ±0.1 inch, or roughly 2.5mm, accuracy. |
In the past, unclear calls could disrupt the flow of a match, create unfair outcomes, or hand unexpected luck to one side. No matter how skilled a referee may be, it is impossible to follow every player and the ball across a wide pitch for the full 90 minutes without missing anything. With more semi-automated systems and faster VAR results, recent World Cup matches feel both fairer and less interrupted. Since one of the strongest objections to automated officiating is that it can break the flow of the game, the technologies introduced in recent World Cups—and the more precise technologies still to come, including computer vision—will almost certainly make matches fairer and more enjoyable.
Related posts
Why Did Starbucks Give Up on Computer Vision Inventory Management?
Jun 9, 2026
Starbucks pulled NomadGo’s computer-vision-based inventory management system just nine months after rollout. This case highlights the limits of marketing claims like 99% accuracy, the constraints of FOV and on-device AI, and the difficulty of simply adding vision technology onto existing store operations.
Between InGaAs and Si: SWIR Sensor Supply-Chain Risks and the Practical Alternative of NIR CMOS
Jun 20, 2026
As supply-chain risks grow around InP and InGaAs-based materials, not every SWIR application necessarily requires an InGaAs sensor. Below 1,100 nm, silicon-based BSI CMOS can be a practical alternative in terms of cost and supply stability. This article looks at the technical boundary between InGaAs and silicon.
Connecting the Arducam Quad 12MP (IMX477) Color Camera
Dec 29, 2025
An overview of the Arducam Quad 12MP (IMX477) color camera and how to connect it.