In the beginning of May, the 68th Eurovision Song Contest (ESC) took place at the Malmö Arena in Sweden. Video surveillance solutions from Axis were installed throughout the event to support security. At the same time, we were able to test our solutions in a real environment to gain valuable knowledge in terms of test data, user input and feedback. We spoke with the team at Axis about how they’ve used this event as a unique opportunity.
Continued innovation is at the heart of what we do at Axis. One way we achieve this is by working closely with our partners and end-customers to understand how they use our solutions and how they can be optimized and improved in the future.
Working alongside the security team at the Malmö Arena during the Eurovision Song Contest (ESC) has been a prime example of how we bring this to life. Our relationship here can be traced back to 2013, when the ESC was last held in Malmö, Sweden. At that time, the installation included just twelve cameras, and the only analytics was motion detection.
A win-win relationship
When Malmö was once again chosen as the venue for ESC 2024, Peter Narbe, Head of Security for the event and the project group at Malmö Arena, contacted the team at Axis to look for support in implementing a state-of-the-art video surveillance solution for the event, covering both the internal and external areas of the arena.
Commenting on the ESC project, Peter Narbe says: “Our security team at the Malmö Arena are familiar with using Axis solutions, having collaborated previously. We knew what to expect – a collaborative relationship and a high-quality security solution – and it made sense to work with the Axis team once again for ESC 2024.”
While this wasn’t the first time our solutions had been used in this context, having previously been installed at Malmö Arena for sports events, the internationally broadcast event brought new opportunities for our product development team at Axis to test solutions in a real-world environment. We could take real-time data, feedback, and learnings from this temporary installation of our security solutions, while the ESC and Malmö Arena would benefit from enhanced, comprehensive security measures around the event space.
Ted Hartzell, Expert Engineer at Axis, notes this presented an opportunity for the Axis team to learn. “We have not only been able to test new products and solutions, but to gather data that can help optimize our existing products and in the development of new ones,” he says.
“Having worked with Axis previously, we already had the existing expertise to do most of the solution configuration and advanced rules for the event ourselves,” Narbe notes. “This has been a real benefit, as it eased the installation and integration process, making it as seamless as possible. In addition to this, the Axis UX team and product development team worked closely with us in the lead up to the event, ensuring the solution met our security requirements and we were set up for success.”
The core security solution
The security team at Malmö Arena required additional, robust video surveillance technology during the ESC, as existing video surveillance at the venue was only indoors. As Hartzell explains, “the areas outside are public and therefore surveillance isn’t permitted.” However, for the purposes of the event, a significant area outside the arena was cordoned off with fencing, so the required area for surveillance extended beyond the arena itself.
“This is where we came into play”, Hartzell continues. “The fencing effectively made this area private for the duration of the event, allowed us to install a new, temporary video surveillance solution outside, which were be taken down after the event.”
The core event-based security solution was designed using AXIS Site Designer. It made use of more than 40 Axis cameras (panoramic, multi-sensor, thermal, and pan-tilt-zoom (PTZ)) covering both the internal and – predominantly – external areas of the arena. Perimeter protection was a core priority, particularly around areas such as the back-up power generators. Thermal cameras and AXIS Perimeter Defender were used to support this, ensuring the perimeter was kept secure. License plate recognition was used to notify guards when vehicles arrived, and Smart Search analytics was available through the implementation of the latest version of AXIS Camera Station (ACS).
“The Axis solution was particularly valuable,” Narbe comments. “It allowed operators to receive event notifications in real-time 24/7, while also providing them with the option to review and find footage related to specific events that might have taken place.”
A test bed for innovation
At Axis, we have used the project in a number of ways to gather feedback and data for existing solutions, bringing this back to the team at Lund for analysis and, were beneficial, immediate action.
Observing solution use
This has offered a rare opportunity for the product development team at Axis to gather live, real-world feedback from end users on their experiences of using AXIS Camera Station. Malin Cronquist, Project Manager, ACS Configuration and UX, at Axis: “Through this project, we have been able to get some invaluable feedback on areas that can be improved, and developed an understanding about how the solution ought to be configured and used.”
Through this project, we have been able to get some invaluable feedback on areas that can be improved, and developed an understanding about how the solution ought to be configured and used.
With on-site observations, learnings were shared promptly with the team at Axis, as well as being delivered through internal feedback and information sessions. “We were able to give general information and feedback during our initial on-site visit, which proved to be valuable. It meant that if an issue arose, we could troubleshoot it right away. The wider team loved to be a part of this, as it allowed them to be agile and feel they were able to contribute to a world-class event,” says Hanna Samuelsson, Senior Usability Engineer, ACS Configuration and UX, at Axis. “We wanted to present our learnings at an early stage, to share the excitement and pride of our products being used for a high-profile event like ESC,” she adds.
It was taken into consideration as to whether the feedback was only relevant to the specific implementation at the Malmö Arena, or whether it was something that could lead to a development or enhancement in solutions for more customers. Samuelsson explains the rationale for this, noting: “our vision is to create a general solution so that all customers can benefit”.
And this is certainly reflected in the work at Axis. From previous testing, the team has been able to make improvements on how users can configure solutions for larger systems and are keen to draw on new learnings from this latest project with the ESC for the future.
Training AI models
The ESC has provided an almost unique opportunity to gather data that can be used to train our deep learning algorithms. Anthony Hawkins, Deep Learning and Analytics Engineer at Axis, explains how the product development team used the event to test their latest AI applications: “the bottom line is with AI, the algorithm is only as good as the data it trains on.”
He notes that while the team has plenty of existing footage of perimeters, fences, motorways, and cities, they don’t have as much on big sporting and music events, like Eurovision. “Getting new data is wonderful for us,” he continues. “It’s a great opportunity for us to test how effective our algorithms are.”
Getting new data is wonderful for us, it’s a great opportunity for us to test how effective our algorithms are
With up to 9,000 people attending each show at the ESC, and nine different shows taking place, there was a real scale which provided a novel testing ground for analytics, such as people counting and head counting in a crowd dynamic. Not only this, but the nature of this event in particular meant there was also a real variety of clothing, equipment, vehicles, as well as changing light conditions. All of this is integral for testing the analytics for accuracy under challenging conditions.
This also enabled a large variety of data to be captured – something that otherwise would be quite difficult to achieve. Gathering data in this environment meant the team at Axis could do so in an ethical way, in line with all relevant data security regulations.
When it comes to training AI models, Hawkins explains that forty members of the team would review the footage captured and annotate images, identifying people, objects, vehicles, colors, and so on, which then act as the basis for the model training. “In a nutshell, to train our models we go from huge chunks of video to hundreds of annotated frames, and then these are input into our models,” he notes.
Concluding thoughts
This project with ESC has been a continuation of an existing relationship with the security team at the Malmö Arena in addition to several other organizations and stakeholders, and we’re keen to see the further opportunities it presents us to collaborate in the future.
It has also served as a reminder that innovation is about more than thinking differently. For us at Axis, it’s about being agile and continually looking for ways to optimize our solutions. It’s about learning. This has been a unique opportunity for us to learn and act as one, tightly integrated team to test, innovate, and optimize our solutions.