The future is now

George Richardson, co-founder and CEO of AeroCloud, explains how computer vision can transform airport operations and the passenger experience.

The removal of COVID-19 restrictions has allowed a steady return to air travel. According to the UN Aviation Agency, the number of air passengers carried in 2022 was an estimated 47% up on the previous year and is expected to reach close to pre-pandemic levels this year.

But while this is undoubtedly positive news, the industry faces a series of ongoing challenges. Airports are under pressure to accommodate this growing number of travellers and deliver the best possible customer experience, while they contend with issues including staff shortages and ever- tighter operating budgets.

In an effort to overcome these challenges, airports are moving away from the legacy technology on which they’ve relied for years, and embracing developments like cloud, data analytics and artificial intelligence (AI), the latter of which includes machine learning (ML) and perhaps the most exciting of all, computer vision.

But, first, what is computer vision?

Computer vision is a subset of AI. On a simple level, it’s worth comparing with human vision. Humans make decisions based on visual input received from their eyes. If, for example, a person wants to recycle, they can split out their glass bottles, paper and metal from items that can’t be recycled.

It’s a decision – and subsequent action – entirely based on visual input. In the same way, software can be built to allow a computer to make similar decisions based on input received from a camera; so, it can detect items from a recycling conveyor belt visual feed and pick out materials fit for recycling.

Tracking passenger flow – accurately – for the first time

When used within airports, computer vision has the potential to significantly improve the way in which they are managed, helping to accurately predict the flow of travellers, and enabling operational leaders to make confident decisions on everything from resource allocation to commercial opportunities.

Understanding how passengers move through an airport, from kerb to gate, is a decades-old issue. And a hold-up at any part of that journey can have a knock-on effect; potentially leading to long queues, delayed flights, and unhappy passengers.

Airports have trialled various technologies to solve this issue, including WiFi and Bluetooth. Computer vision, though, is the first technology capable of intelligently, accurately, and anonymously tracking passengers throughout their whole journey in real-time – from arriving at the airport, through security and departure halls to the gate, and then boarding the aircraft.

In essence, computer vision takes visual inputs from CCTV cameras located throughout the terminal and makes decisions based on that data. Combining this visual data with AI and ML algorithms allows airport operators to spot trends, draw learnings, and predict future scenarios to inform more accurate decision-making.

Across most sectors, including airports, there have naturally been worries about AI from consumers when it comes to privacy and security. Importantly, computer vision can be used anonymously by airports while being fully compliant with all local privacy legislation, such as GDPR in the European Union and the UK. No data is stored, and passengers can’t be identified as individuals.

What can airports achieve with computer vision?

By monitoring passenger flow in real-time, alerts can be triggered in response to any unexpected bottlenecks that may occur, meaning they can be quickly and effectively addressed with additional resources.

For instance, staff can be notified when passengers have been waiting at a check-in desk for more than 20 minutes; a situation that can be quickly remedied by opening additional desks.

Facilitating smoother passenger flow is a key objective for airport operators. But the benefits of using computer vision go beyond this one use case.

Computer vision can enable better resource planning. By connecting the data insights, it provides with AI and ML technologies, airports can maximise their – often stretched – resources, by redistributing employees to where they’re most needed based on real-time or predicted operational needs.

If they know, for example, that departure halls are especially busy between 3pm and 5pm every Friday, they may consider emptying the bins more regularly or scheduling additional maintenance checks on the bathroom facilities.

The technology can be used to maximise the profits of concessions, too, as getting passengers quickly through security and into the departure hall is a key source of revenue for many airports in Europe.

Airports can combine insights from computer vision with data from other sources – like a business intelligence platform – to identify valuable trends data. This can be passed on to inform planning for concession partners, such as duty free, retailers, and hospitality providers, to maximise customer spending.

Should there be a large number of passengers arriving at a particular time each day, concessions could consider running promotions or creating a pop-up to encourage sales, for example.

Indeed, providing adequate retail and hospitality opportunities, being certain the correct number of staff is available to ensure facilities are kept in good condition, and limiting the amount of time people spend queuing will all help deliver a better experience for passengers.

This, in turn, can engender greater loyalty among travellers and encourage them to return to a given airport should alternative options exist.

Tried and tested

It may once have been a science fiction writer’s dream, but today, computer vision is very much a tried and tested reality.

Liverpool John Lennon Airport, for example, uses it. Historically, the UK gateway used various systems and manual methods like queue counting to track how individuals travelled through the building, but none delivered the desired insight. But by identifying peaks and bottlenecks, and where queues were most likely to be, computer vision enabled its operations team to flatten planning curves and increase efficiency.

In addition, the accurate information it provided helped inform its marketing team’s decision-making and improve spend per head in its retail and concessions environments.

And, on the other side of the Atlantic, Florida’s Sarasota Bradenton International Airport used the technology to automate gate allocation and passenger prediction analytics, helping it take on new airlines and increase its annual passenger figures. Going from 900,000 passengers in 2019 to over 3.4 million in three years, it was the US’s fastest-growing airport of 2022.

Overcoming challenges

Airports today face a variety of challenges. While numbers may be returning after the pandemic, many are still struggling with the fall in revenue and debts incurred during the peak of COVID, not to mention the shortage of employees.

Computer vision overcomes challenges such as these. It uses existing CCTV infrastructure, which eliminates the need for expensive replacement equipment, and as no camera images are stored or accessed, it complies with data privacy regulations. Most importantly, it’s proven to increase efficiency and reduce friction in flow, improving the passenger experience while delivering additional benefits around security, marketing, revenue, and more.

This is a new technology for airports. These current applications are only scratching the surface. For the airport sector, when it comes to using computer vision, the sky really is the limit.