Making the most of powerful AI video surveillance

6/10/22, 9:30 am

How machine learning finds needles in a data haystack to protect passengers and infrastructure

Smart cities rely on complex transport systems to get people where they want to go quickly, seamlessly, and safely. These systems require round-the-clock, intelligent real-time monitoring to ensure smooth operations and protect against physical and cyberattacks.

Artificial intelligence (AI) can make it so. Using machine learning (ML), an AI model is trained using the vast amount of data and video collected from existing real-time and video surveillance systems.


Putting surveillance on steroids

Video surveillance can sometimes be controversial but is generally widely accepted as a pillar of safety in public places. The problem with surveillance on such a large scale? The sheer volume of data from video systems makes it impossible for operations staff to effectively monitor – or act on – vision from all cameras.

It's a problem beyond a scalable human solution; it needs the power of AI – specifically, machine learning (ML). ML is the science of generating computer programs from data to perform tasks without being explicitly programmed to do so. ML powers language, facial recognition, biomedical processes, and financial services.


ML as a transport safety tool

ML capabilities supplement standard video surveillance systems – allowing operators to be responsive to safety issues.

Imagine security guards at a sporting event watching hundreds of camera feeds. The trained model can monitor and spot anomalies much faster than humans and at a significantly higher volume. For example, an AI model trained by ML could find violence as it starts to occur and create an alert without missing the action on the other feeds.

Whatever the ‘out of the ordinary’ behaviour or event, ML detects it and generates an automated alert – fast, so more cameras are actively monitored by fewer staff.


How ML manages the new risks of autonomous vehicles

Once a far-off fantasy, autonomous vehicles are a rapidly evolving reality for personal and public use. Capable of sensing their environment and operating without human intervention, they enhance traffic safety. But they also introduce new risks:

  • threats to personal safety if the car operates without a supervising driver
  • vehicle hacking and malfunctions
  • responding to complex driving conditions
  • providing automated safety surveillance on driverless transport.

Without human oversight, AI video surveillance powered by ML is critical to ensuring autonomous vehicles are literally staying in their lane and passengers are arriving safely.


Live monitoring with 5G and edge processing

5G allows for more real-time monitoring of live bus and rail systems – along with 'edge processing' of data. Edge processing involves acting on data as close to the source as possible – rather than in a central, remote data centre. Consider a moving train – even a sophisticated smart train is only as smart as its internet connection. Black spots are inevitable; the time lag could put passengers at risk if you're reliant on off-site data processing. Blending autonomy and connectivity is crucial to get value from these safety systems.


The trust factor

Safety and security are key to people trusting public transport. Surveillance and visible indicators are critical tools – provided they prompt real action like recognising peak demand and deploying more vehicles. AI does best when automating repetitive tasks that rely on identifying patterns and pattern anomalies. Using it to supplement existing processes and tools to improve safety and security is a natural next step to increase trust.

If we ask customers to trust our systems, it's imperative that surveillance processing meets personal data protection compliance and expectations. A sophisticated data management system is required to store, secure and process data through various checkpoints — share that data on a zero trust basis.


How AI helps Lisbon see the transport future

Portugal’s capital, Lisbon, is a leading European smart city. In collaboration with Lisbon City Council, NEC implemented the Lisbon Intelligent Management Platform, a smart platform integrating various data sources with AI.

“The Lisbon Intelligent Management Platform allows us to gather this information and learn from the past so that we can see the future,” said Joao Tremoceiro, Lisbon City Council’s Chief Data Officer.

“Citizens expect much faster response, and many problems faced by modern cities can only be solved by intelligent platforms that integrate data to allow the government to be proactive and for different agencies to collaborate.”

Video surveillance is quickly analysed, and suspicious behaviour is flagged with emergency service providers. The Council has improved public safety by monitoring and managing traffic and transportation systems.


Partnering for smart transport solutions

NEC Australia partners with Public Transport Authorities and Public Transport Operators to provide scalable solutions that optimise operational performance, enhance passenger experience and improve network safety.

Smart Transport Services are changing the way we think, plan and undertake journeys by delivering the latest in machine learning AI video surveillance.

Find the needle in the haystack in your smart city transport system.

Talk to the experts