Choosing a video analytics solution – a guide for cities

August 25, 2021 11:27 am

traffic in Hong Kong at sunset time

traffic in Hong Kong at sunset time

The global video analytics market is expected to increase from US$4.9 billion in 2020 to US$11.7 billion by 2025, with smart cities tipped as a key growth area.

This is being driven by several factors, including an increased focus on public safety and crime reduction, a growing demand for real-time actionable insights from camera data, and the falling costs of video surveillance equipment.

Many cities are now beginning to see the potential of video analytics as systems were deployed during the pandemic to monitor social distancing, mask-wearing, transport usage and more.

With a growing interest in real-time video analytics technology, this guide outlines the criteria for choosing the right solution for your city.

Define city challenges

It’s crucial to start with the challenges to be solved, rather than the technology itself. Define some of the key challenges that the city is facing and prioritise them. Because modern video analytics solutions are multifunctional and growing in capabilities all the time, don’t limit the issues that you list.

For example, the right system could help to manage everything from traffic congestion and road safety to littering, fires and violence. The latest systems use behavioural analytics rather than outdated pixel technology which typically only recognises and classifies static objects. Behavioural analytics can assist with recognising risky behaviour, fighting or traffic violations, and are constantly evolving via machine learning.

Systems thinking

Although it’s important not to be too prescriptive, think about the system performance issues that are most important to you, weighing up considerations such as the false positive rate, recall rate, and hardware footprint, as well as the total cost of ownership (TCO). On budgeting, you’ll need to take into account the cost of a system licence, hardware and ongoing maintenance and support.

Weigh up different financial models. Some providers offer ongoing software licences of one to five years or a software subscription model, which can better allow local governments to balance costs and demonstrate ROI to budget holders.

Think about vendor requirements and due diligence. Both start-ups and established companies have benefits to offer as long as you assess each vendor’s financial stability, vision and roadmap, position in the market and track record.

Privacy is a key consideration for cities and it’s worth noting that the latest video analytics systems don’t require facial recognition or the collection of any personal data.

Research the market

Once you have a solid evaluation of the city’s requirements and vision, you’re in a good position to research the market. An independent consultant or systems integrator can be helpful here if you’re unfamiliar with the sector.

You could either issue a Request for Proposals or take a more open-ended approach through a Request for Innovative Solutions or a challenge-based procurement process.

It’s a good idea to solicit responses from at least three vendors, including a detailed analysis of the system available and interviews. Providers should be open about explaining their system, including the underlying algorithms and privacy and security protections in place. A reluctance around this could be a red flag.

Try before you buy

Before making a long-term investment, a pilot programme is the most effective way to test out the system and get other city stakeholders on board.

For a solid proof of concept, it’s worth applying the video analytics solution to 10 to 15 percent of the potential camera base in the city for a period of four to six weeks. This is the only way to be confident that the product can solve your problem. To simplify the process and allow good comparison, it is advisable to select one vendor for a production pilot and then move to the next if the solution doesn’t meet expectations.

Blog Posts