top of page

Addressing Football Hooliganism with technology

Updated: Jan 18

Football Stadium - Image by Mitch Rosen
Live audience at a football game

Football is the biggest sport in the world. The FIFA World Cup 2022 has just ended, and it was a successful event, with many live spectators and online viewers all across the world.

However, violence and unrest, and stampedes are a common issue with large crowds. It is vital to prevent such incidents from happening when dealing with large amounts of people within an area.

More recently in 2022, there was the case of a riot and stampede that happened in East Java, Indonesia during October 2022, where more than 130 people were killed, and 320 injured according to The Straits Times. The police had fired tear gas in an attempt to control the riots and unrest, which led to the stampede. This is a notable event of one of the worst stadium disasters that has ever happened.

Unrest and violence from crowds has been an ongoing problem for years in football. Events of riots and stampedes have been reported countless times in several countries. Stampedes are a dangerous phenomenon, and may cause many injuries and fatalities. More recently, the Itaewon crowd crush in South Korea during the Halloween season killed 155 people and left 152 people injured. However, according to a study done by crowd monitoring technology expert Dr Anders Johansson, there are patterns of behaviour in stampedes that, if spotted early enough, can prevent the stampede entirely.

Different types of technology can be applied to such a sport to improve operations. In football, technology such as artificial intelligence, video analytics and machine learning are utilized to improve game experiences.

Goal-line technology

One such current application used in football is the Goal-Line technology (GLT). The Goal-Line Technology consists of several types of technologies that work together to monitor the path of the ball, and determine if it has crossed the goal line.

The processes involve data input, data conversion and the use of AI algorithms. For data input, the AI is trained to differentiate between a ‘goal’ and ‘no goal’. Images are added to the machine as part of training. This may train the AI to identify the goal line in adverse weather conditions, and also only count the score when the game ball crosses the goal line. For data conversion, the AI makes the decisions, where the referee may not be able to make precise judgement with the naked eye.

The Artificial Intelligence domains used are machine learning and deep learning in computer vision. Such technology is used to calculate the ball’s coordinates, position and speed to determine if a goal has been scored. On top of that, deep learning with a convolutional neural network is used to recognize the ball in the field to monitor and analyze its path. Additionally, it is also applied to the replay footage captured by cameras within the playing zone.

The benefits of the GLT is that it is extremely accurate and precise, reduces biasness and human error, and is able to be documented for justification purposes. The GLT has helped to aid match officials like referees and umpires in their decision making in matches.

Using technology to address football hooliganism

As mentioned above, it is essential to detect threats before they happen so that casualties are minimized. To better address and mitigate riots and stampedes, a similar, integrated use of facial recognition and video analytics can be used to further improve safety of the game.

How do the devices and machines identify riots and stampedes before they happen? The video analytics technology work through the use of real-time detection. In this step, the machine visualizes and is able to detect events in real time through data analysis and deep learning for event recognition. To solve real data collection issue for deep learning, the idea is to let the system learn with simulation. In this case, the machine learns by analysing the inputs of simulation data. Then, it undergoes training through a framework to learn a classification model to recognize anomaly events of riots in early stages, on top of using crowd detection. Lastly, the accuracy provided by the machine is thoroughly checked for validation.

With this deep learning system, the machine is able to make proper judgement and decisions. It can be used in a computer vision system that can be implemented at the entry point or area of spectators. A facial recognition system can be deployed to identify each and everyone of the spectators at the stadium. Additionally, spectators with a record of misconduct - with previous cases of offence - can be placed on notice or ‘high alert’ to be monitored more closely.

Additionally, linking the person with their identified personal vehicle is possible with a smart parking system in CCTV cameras, which may include blacklisting using technology such as License Plate Recognition (LPR).

During the match of the football game, smart CCTV cameras equipped with the developed convolutional neural network can detect any sort of suggestive violent or hooligan behaviour from the spectators, via video analytics. If the misbehaviour is detected, the security officers on patrol can be notified of the details of the suspect including their face, name, age, race and gender via communication devices such as their watch.

Impacts on consumers and businesses

With this, it brings benefits to both consumers and businesses. Consumers include the sports fans, athletes and stadium goers. For consumers, including sports fans and athletes, their physical safety is protected by preventing stampedes and violent crowd behaviours. Spectators can enjoy a better viewing experience without disruptions and athletes will have an optimal environment to compete. Matches can go on without fans from either team causing too much trouble, which reduces friction in the football community.

How does it benefit businesses? These businesses include stakeholders, event organizers, stadiums, governing bodies and corporates.

Firstly, it reduces the time for detection of such cases. real-time violence detection is useful in such live programmes. The technology can detect them very quickly and easily in the stadium, and it reduces the time needed to mitigate/ solve the issues that can cause a potential hindrance. Thus, they can plan well for and in such events. In turn, this decreases the need for additional manpower. This can enable the stadium to deploy minimal security. The computer vision is also able to work 24/7, which is flexible for any time of day for games. With this, stadiums can also more confidently obtain full capacities at their facilities as a result - this benefits with more revenue and commercialization. Lastly, it will uphold the reputation of the stadium. Large organizations such as big clubs and sponsors would want to have their activities held in such stadiums where they can not only have access to a safe, efficient area, but also a wide reach for commercialization purposes.

Thus, businesses can leverage on such technologies to better improve security in the physical environment of football stadiums, to create a safe and optimal environment, thereby improving overall game experiences for all.

bottom of page