Aggregating, automating and augmenting AI for SG’s smart city
SINGAPORE’S vision of becoming a smart nation to create an economically competitive global city and a livable home for the future is fast becoming a reality.
Indeed, a recent global smart city ranking put Singapore in the first place, ahead of major cities such as London, New York, and Seoul.
To edge closer to this dream, Artificial Intelligence (AI) will be harnessed to a much greater extent, working with sources of potentially useful data, such as networked video cameras, sensors and Internet of Things (IoT) devices as they are deployed all over the city.
AI and video technology will be of particular importance.
Today’s processing power means computers can analyze huge volumes of video data simultaneously, recognizing objects, people, and even behavior.
Some applications in a smart city include the use of videos for city and traffic planning, or even license plate or facial recognition to track criminals. These potential applications for this advanced type of capability will change the way we live and work.
Here are three steps to harness the power of video to quicken Singapore’s pace as a smart nation:
Aggregating video sources
For a city to be truly smart, its sources of information need to be first connected in order to provide the most complete picture possible. In Singapore, this is still a work-in-progress.
An aggregated network of video cameras, for instance, provides a vast amount of information. When connected to AI, the system can then provide insights to regulate traffic or find a missing person.
Currently, GovTech is developing a nationwide sensor network called the Smart Nation Sensor Platform.
This platform is aimed to provide the communication backbone to facilitate smart metering in residential areas, help with public transport planning and enable the creation of innovative solutions and applications via smart street lampposts.
Aggregated networks of video sources should also be connected to an open-platform video management system (VMS).
This will allow the network to be expanded easily, without having to use one particular kind of device or camera, meaning authorities can use whatever solutions best fit the task.
As well as making the addition of hardware easier, it facilitates the creation of even bigger datasets, in order to provide a more accurate picture in future.
Automating insights for safer cities
With systems in smart cities being connected all around, the massive amounts of video and sensor streams coming in mean that we need to use the data in more clever ways.
However, this progress has been impeded until recently as the computing power needed to process and analyze those streams has simply not been available.
Automation will be at the center of the video management system as the sheer size of data sets would be impossible for human beings to utilize.
This takes much of the tedious burden of manual security scanning away from security guards.
Today’s processing power means AI can be trained to recognize behavior, not just objects.
For instance, if a person is loitering suspiciously, this can be identified as unusual behavior, then an alert can be sent to a security officer to investigate on the situation. This capability will reduce necessary human intervention and ease attention fatigue for security operations.
In a smart city scenario, we will solve public safety and operational challenges in ways that are truly impactful. For example, police cars in Singapore are currently fitted with about 8 – 10 cameras each.
When looking out for blacklisted car number plates, the police are able to minimize the deployment of manpower by leveraging license plate recognition instead.
Furthermore, with the new security Industry Transformation Map (ITM) underway, more security companies will move away from just supplying manpower and focus on the use of technology to deliver security solutions effectively.
This is especially so in the backdrop of a manpower crunch in the security industry.
However, the real transformation to a smart city does not simply stop at the automation level, instead of going on to amplify human intelligence.
Augmenting human intelligence for a truly smart nation
As the city continues to grow, and hardware acceleration such as ultra-high definition cameras allow us to have all of that aggregated video and sensor information, more will be used to analyze, assist and augment the way we manage buildings, incidents, and people.
However, the use of video is not limited to security applications.
The potential of video technology remains relatively untapped. Retail, for example, can make use of facial recognition software to enhance the customer experience.
For instance, loyalty programmes can be activated when facial recognition recognizes a returning customer.
When connected to a business’s point-of-sale system, sales staff can even recommend items the customer might like based on previous transactions.
This is just the start. AI is going to provide the basis for the many different applications which will help Singapore earn its status as the world’s first truly smart nation.
Contributed by Benjamin Low, Vice President, APAC, Milestone Systems.