What does a smart city look like? How video surveillance AI is changing our cities

Writing for IFSEC Global, Lucy Booker explores just how AI-based video surveillance is changing urban environments – and what future smart cities might look like.

‘Smart cities’ – also known as ‘safe cities’ – are on the rise around the globe. These hyper-connected infrastructures collect data from buildings, people and devices to enhance our daily lives.

Whether it’s optimising waste management, air quality or traffic flow, video surveillance Artificial Intelligence (AI) systems play a significant role in optimising our future built environment.

But how is smart video surveillance being used in cities now? And in particular, how does Automatic Number Plate Recognition (ANPR) – the most common use of AI-based video surveillance – play a significant role?

---

How will our cities change?

According to the UN, two out of every three people will live in cities by 2050 – which equates to an additional 2.5 billion people. Our urban centres are evolving and there’s no doubt that this kind of population density needs management.

The solution? Smart cities: a living organism defined by the European Commission as “a place where traditional networks and services are made more efficient with the use of digital solutions for the benefit of its inhabitants and business.”

This level of connectivity means breaking down “data silos” in physical security is crucial. So, it’s no surprise that video surveillance AI that integrates with other solutions – all while tackling multiple challenges – is being presented as the future.

“When it comes to enhancing public safety and maintaining security, having a complete picture of your environment is key”, said Jean-Philippe Deby, Business Development Director at Genetec. “Having a modern unified platform that encompasses systems like video surveillance, access control, ANPR, intrusion and analytics goes some way towards achieving this.”

The most prolific use of video surveillance technology in smart cities involves ANPR. So, let’s look at how it’s used…

How is AI-powered ANPR changing?

According to Omdia’s Smart Cities Database, the global market for ANPR products exceeded $1.4bn in 2021. Much of this growth has been driven by new wireless, battery-powered and solar-powered hardware – with market demand for rapid deployment systems up by over 18% in 2022.

The accuracy of these systems is increasingly reliable, as Bryan Montany, Senior Analyst at Omdia explains: “Analytics relating to Optical Character Recognition (OCR) processing have improved the accuracy read numbers from vanity plates, as well as the ability to identify these numbers at night and in adverse weather conditions. ANPR software has also become more capable of accurately identifying the makes, models and other distinguishing features.”

As ANPR gets more sophisticated, the hardware itself is getting smaller. For example, US-based firm Supervision claims that its Picopak mobile technology is the “world’s smallest LPR camera.” Measuring just 10 x 10 x 4cm, the device is so compact it can be used on a scooter, or as the website shows – a child’s car.

How is AI-powered ANPR being used in UK cities now?

In practice, UK cities such as Bath, Birmingham, Nottingham and Oxford are already using smart ANPR to tackle issues such as pollution.

In Bristol, local company SEA is providing 49 of the firm’s ROADflow Fusion cameras to detect vehicles that don’t meet the zone’s emission standards. The system also links with third-party solutions, such as payment providers.

More broadly, multi-functional spaces are also appearing at transport hubs. For example, parking operator APCOA has announced the first of its seven Urban Mobility Hubs in Carmarthen, Wales. These points will include electric vehicle charging points facilitated by ANPR technology, alongside other services such as Wi-Fi-pods for remote working.

Improving metropolitan transport systems will remain a focus for expanding cities. According to Emergen Research, traffic management accounted for the largest revenue share of the global ANPR system market in 2020.

Changing drivers’ behaviour by putting a price on various factors, such as car emissions, peak driving times and economic profile – as seen in products such as Kapsch’s digital traffic management system – attempts to incentivise road users into changing their behaviours.

“Understanding broader traffic trends”

Looking at the technology itself, video surveillance AI is mirroring some of ANPR’s capabilities, as Bryan Montany outlines: “It’s now possible for conventional cameras without optical character recognition (OCR) capabilities to accurately determine the make and model of passing automobiles.

“But, in a traffic monitoring context, metropolitan agencies responsible for smart cities aren’t concerned with catching individual offenders and instead are eager to understand broader traffic trends. In those cases, network cameras can perform similarly to ANPR equipment – but cost hundreds of dollars less per camera.”

If the distinctions between ANPR and video surveillance AI are less obvious, how are they working together in smart cities now?

How is AI video surveillance and ANPR being used in smart cities around the world?

It’s estimated that China has over half of the world’s smart cities. The emergence of edge computing, cameras and sensors using 5G wireless connections is expected to accelerate the use of AI-powered video surveillance here – and around the globe.

In Shenzhen, hailed as the ‘world’s first entirely smart city’, sensors and cameras have almost the whole city under constant monitoring and surveillance. Data is collected, centralised, and integrated to inform urban management, planning and individual decisions.

Alongside China, other countries such as the UAE, Saudia Arabia and Indonesia, are expected to leverage video surveillance AI to its fullest potential, as Bryan Montany explains: “Over the next five to 10 years, the use of cameras will expand beyond capturing data related to automobiles and begin to capture the biometric and behavioural characteristics of drivers. Instead of just monitoring vehicles, these cameras will monitor the people behind the wheel.”

In Western cities, video analytics systems are proving their use with multiple applications. The Belgian city of Genk, for example, is using Nokia’s Scene Analytics to measure noise levels on a commercial street and report back to concerned residents. Further afield, the same technology was trialled in Melbourne, Australia, to help authorities combat illegal waste dumping.

However, privacy concerns are major considerations for many countries. In the UK’s Biometrics and Surveillance Camera Commissioner’s Annual Report 2021-2022, the commissioner Professor Fraser Sampson cited ANPR as an area of focus and said of the UK: “Now more than ever, we need a clear, comprehensive and coherent framework to ensure proper regulation and accountability in these crucial areas.”

How is AI video surveillance helping law enforcement?

One of the biggest challenges for smart video technology is keeping up with an ever-expanding number of urban threats. As a case in point, security software developer Ambient AI recently added 25 new behaviours to its library – bringing the total number of scenarios it monitors to more than 150 – including weapon detection, crowd commotion and unauthorised entries.

Retrospectively, video surveillance AI also allows for an in-depth review of footage in smart cities to find certain people or license plates. For example, heat maps can help pinpoint which property on a road might be a centre for criminal activity. It can also be used at the entrance to public buildings, such as schools or government buildings, to measure footfall.

Furthermore, in the event of a public health emergency, such as the COVID-19 pandemic, data can be gathered via AI-powered video surveillance on whether people are wearing face masks or following social distancing rules.

What will a future smart city be like?

As ever, what ‘can’ be achieved must be carefully balanced with what ‘should’ be achieved, as privacy concerns remain ever-present – albeit more so in some areas than others. Yet, there is a clear role for video surveillance to play, as ever-more data is generated, collected and used to create these ‘hyper-connected infrastructures’.

So, if you’re driving into a smart city in 10 years’ time, you just might be rerouted around an incident, alerted if your car isn’t allowed within a pollution-controlled zone and automatically guided to the most convenient parking space.

A hassle-free journey in a safer environment might just be within reach – all thanks to smart cities and the AI-powered video surveillance technology that underpins them.