Stadium security: between overview and identification

Shortly before kick-off, a football stadium becomes a highly congested security zone. Tens of thousands of people move through entrances and concourses simultaneously; supporter stands fill up; queues form; and traffic flows become congested on access roads and in car parks. For the security organisation, one thing matters above all else during this phase: maintaining an overview.

However, an overview is not the same as identification. A camera that reliably captures a crowd fulfils a different purpose to a system designed to identify an individual person in the event of an incident. It is precisely this distinction that lies at the heart of the guide ‘Video Technology & Data Protection in Stadiums’, developed by the German Football Association (DFB) with the support of the ZVEI.
For the revised edition of this practical guide – first published in 2013 – technological developments were taken into account alongside changes to the legal and regulatory framework. The result is less a product recommendation than a planning concept: cameras, image quality, storage, control centres, security organisation and data protection are viewed as an integrated system.

Not every view needs a face

The underlying principle is remarkably topical: Modern video technology should enable a high level of security without having to continuously record all spectator and operational areas at a quality that allows for individual identification. At the same time, the systems should support crowd management, the identification of offenders and cooperation with the security authorities.

The guide resolves this apparent contradiction by distinguishing between different surveillance objectives. For a general overview, the focus is initially on detection. Where an incident needs to be monitored more closely or a person identified, the required level of detail increases. At particularly sensitive points, additional camera angles may be added. The three-stage model described accordingly ranges from detection, through detection and identification, to recording from multiple angles.

The planning question is therefore no longer: ‘What is the camera’s maximum resolution?’, but rather: ‘What information is required at which location for which security process?’

When pixels alone are not enough

DIN EN 62676 provides technical guidelines for this. Among other things, the guide specifies 25 pixels per metre for ‘detection’, 125 pixels for ‘recognition’, 250 pixels for ‘identification’ and 1,000 pixels per metre for ‘verification’.

In practice, however, the actual planning work only begins with these figures. A stadium entrance requires a different image quality to an entire supporters’ stand. Corridors and access points pose different challenges to car parks or the area surrounding the stadium. And even a high theoretical resolution is of little use if backlighting, unfavourable viewing angles or motion blur render the crucial details unusable.

The technical appendix therefore lists camera position, lighting, dynamic range, depth of field, exposure time, frame rate, compression and weather as factors influencing identifiability. The guide distils this relationship down to a principle that is as simple as it is important: video technology must be understood as a system.

A stadium, in particular, demonstrates these interdependencies impressively. Pyrotechnics create extreme lighting conditions within seconds. People move quickly. Stands require large depths of field, and cameras sometimes operate over considerable distances. A system must therefore prove its worth not under laboratory conditions, but under floodlights, backlighting, smoke, movement and changing weather conditions.

From the camera image to the security process

The role of video technology is changing even more fundamentally where it is not merely required after an incident has occurred.

The DFB and ZVEI classify its use into four phases: prevention, protection, response and documentation. Video systems can detect unauthorised access or vandalism, provide situational information for emergency services and enable events to be traced forensically at a later date.

The operational benefits are particularly evident in crowd management. Movement and queuing patterns can be assessed, critical crowding identified and escape routes monitored. Added to this are traffic control, car park management, number plate recognition and the rapid provision of images to emergency services.

In this way, the camera image evolves from a means of documentation into operational intelligence. A security control centre does not need to display as many images as possible. It must extract, in a timely manner, the information from the available data that can be used as a basis for action.

It is precisely at this point that economic added value is created. If it is detected early on that an unusually high density of visitors is forming at a particular entrance, staff can be redeployed. Traffic and visitor flows can be managed more effectively, disruptions assessed earlier, and security personnel deployed in a way that better meets requirements. The guide explicitly cites more efficient staff deployment, as well as learning, simulation and process improvement, as potential benefits of modern video technology.

Security involves ongoing costs

However, the economic perspective must not end with the purchase of the cameras. A professional stadium solution includes network and storage infrastructure, control centre workstations, software, interfaces and cybersecurity. Depending on the existing infrastructure, structural measures may also be required.
Added to this are running costs. Maintenance is required to ensure the agreed technical functionality is maintained, whilst services are needed to adapt systems to technological progress. The guide explicitly mentions updates, upgrades and ensuring the appropriate level of IT security. Training is equally important to ensure that operators can handle the technology even under time pressure. For temporary requirements, even hosting, ‘Video as a Service’ and ‘Pay per Use’ are cited as possible models.

This is particularly interesting for stadiums, as requirements are not identical for every event. A regular league match presents a different risk profile to a high-risk match, an international final or a concert. Additional temporary technology can therefore be more cost-effective than permanently maintained infrastructure for rare peak demands.
In professional football, responsibility for the proper conduct of a home match generally lies with the host club or organiser. This includes providing adequate security staff and the necessary safety measures. The associated costs fall within the organiser’s remit.

The bill does not end at the stadium gates

Outside the immediate event area, there is a second cost factor: public safety.
Police officers secure railway stations and access routes, separate rival groups of supporters, escort fans on their journeys to and from the venue, and respond to disturbances in public spaces. High-risk matches, in particular, can entail considerable additional effort.
The extent to which this effort can grow is illustrated by the long-standing debate over charges for additional police operations at particularly high-risk large-scale events. Case law has now generally confirmed that corresponding fee structures under state law may be permissible under certain conditions.
This does not imply a general nationwide obligation for football clubs to bear these costs. However, the debate highlights that the economic considerations surrounding stadium security have long since extended beyond private security services. Clubs and organisers invest in their own security organisations and technology, whilst federal states and local authorities simultaneously allocate considerable resources to security in public spaces.
Added to this are differing ownership and operational structures. Stadiums may be owned by clubs, private operating companies or local authorities. Investments in server rooms, networks, control centres or other security-related infrastructure may therefore, depending on the model, directly or indirectly affect public budgets.

Video technology will not eliminate police costs. However, better access to information can help to assess incidents more quickly, deploy resources in a more targeted manner and support cooperation between stadium management, security services and the police.

Data protection begins before recording

The more capable a video system becomes, the more important it is to limit its use. This is because technically feasible surveillance and legally permissible processing are not the same thing.

Images of identifiable individuals constitute personal data. For private clubs and stadium operators, a balancing of interests is required in areas accessible to the public. A legitimate security interest alone is not sufficient; its necessity must be weighed against the rights of the data subjects. If an external security service is deployed, the responsibilities under data protection law must also be regulated accordingly.
Transparency requirements are also significantly higher than in the past. A camera pictogram alone is not sufficient. Visitors must, amongst other things, be able to identify who is responsible for the processing, the purpose of the video surveillance, the legal basis on which it is carried out, and how long the data will be stored. Further information must be made available.
The preliminary concept has far-reaching implications for operational organisation. Camera locations and surveillance areas must be documented, access rights defined and responsibilities established. As a general rule, a short retention period is envisaged for video recordings. In the event of security-related incidents, deletion may be suspended; longer retention periods must be justified. In the case of extensive and systematic surveillance of publicly accessible areas, the data protection impact assessment must also be taken into account.

Who owns the view?

The question of responsibility becomes particularly interesting when a club, security staff and the police use the same technical infrastructure.
If the police themselves take control of the cameras and analyse the footage for their own purposes, they assume their own responsibility under data protection law. Who physically owns the camera does not, therefore, automatically determine who is responsible for the data collected by it. Agreements should therefore clearly stipulate who is permitted to access which parts of the system, when and from which workstation.

What may initially appear to be a mere administrative detail is of fundamental importance for networked security architectures: technical integration must not obscure legal responsibility.

Some images should not be captured at all

This approach is most consistently applied where data protection is directly incorporated into the technology.

The guidelines recommend using so-called privacy filters wherever available. These can be used to initially obscure faces or areas, which are only made accessible when there is a specific need for identification. Private property outside the stadium grounds must not be recorded. Toilets and changing rooms, as highly private areas of personal life, must be completely excluded from video recordings. Where a camera pan could theoretically reach such areas, technical masking should prevent any usable image data from being generated in the first place. This makes data protection a planning factor. It is not merely the question of when an image is deleted that is relevant. What is decisive is whether this image needs to be captured at all.

The stadium as a model case

It is precisely in this respect that the significance of the guidelines extends beyond football. Airports, railway stations, event venues and other high-traffic facilities face comparable challenges: large crowds, changing risk scenarios, private operators, public security agencies and growing technical capabilities.
The stadium serves as a prime example of how improved video security does not necessarily mean more surveillance. Rather, it requires a tiered approach to information quality: an overview where an overview suffices; identification where it is reasonably necessary;
multiple perspectives where a specific high-risk area demands them; and deliberate invisibility where no legitimate security purpose justifies recording.

At the same time, it is clear that security is always an economic issue as well. Clubs, operators and the public sector each bear different aspects of a security architecture whose costs cannot be reduced to just cameras and servers. Maintenance, staff, security services, the police, data protection and public infrastructure must also be factored in.
The real strength of the DFB-ZVEI guidelines therefore does not lie in any single technical recommendation. Rather, it lies in the stadium being viewed as a security process in which technology, organisation, law and economics interact.
Ultimately, it is not a question of how much a video system can see. What matters is what it needs to see, who requires this information, the resources it consumes – and where the technology’s field of view should deliberately end.