09 August 2021


PSI interview with Pauline Norstrom

This month Simon Banks talks to Pauline Norstrom, CEO of Anekanta Consulting, an AI innovation and strategic advisory company working with international defence, security businesses and stakeholders. Pauline was previously BSIA Chair, is now an Honorary member advising the board. Pauline is considered an expert on all things CCTV and AI in Video Surveillance.

What role is video playing in the current COVID world and what opportunities are there for Installers?

The exodus from the office environment caused an increase in empty and partially occupied buildings, increasing vulnerabilities for all sorts of crime. Video plays an important role in providing situational awareness remotely and informs a fast and effective response. We anticipate that there will be an increase in camera deployments on otherwise alarm-only buildings, thus enabling security monitoring to become more proactive around the perimeter and to play a vital role in public safety.

Video is most valuable when installers combine it with other sensor data such as access control and alarm events. Installers should consider seizing the opportunity to offer video verified alarms given that both edge sensor and cloud technology are so easy to access, especially when underpinned by robust and cyber secure communications technology. By adding cloud-based video services, the installer can create increased value for the user and generate extra recurring revenue for themselves. Adding this functionality, with minimal disruption, also increases up-time and lowers the total cost of ownership.

Video is playing an increasingly important role in the COVID world by making spaces contactless, such as automated facial recognition software used for access control purposes. In retail and hospitality, video-based occupancy monitoring has become commonplace in managing COVID risks. The technology can reduce the manual burden of counting in and out the number of shoppers in a store.

It is my belief that COVID has caused an almost “wartime-like” acceleration in the development and acceptance of video and cloud-based technologies. We also need to accept that COVID is probably here to stay, therefore the need for video combined with other sensors to create safe, contactless environments can only increase.

What differences can concepts like AI (Artificial Intelligence) and DLVA (Deep Learning Video Analytics) make in terms of CCTV?

AI has really come into its own in physical security technology over the last 12-24 months. But this did not happen overnight. Many tries and fails over many decades and billions of dollars of investments led to the emergence of usable technology. The early attempts to recognise objects in images were unreliable and required weeks and months of fine tuning to remove unwanted alerts.  

These early technologies relied upon comparison images which were manually categorised and labelled. The AI would generate its best attempt at finding a match in the image with varying degrees of success. The application of AI has matured and evolved since then, as has the processing technology and image quality leading to the advent of deep learning networks which improve their results over time.

Deep learning AI is now deployed within cameras at the edge to detect or track an object of interest and discard non-useful information. The cameras can also operate as a collective intelligence, providing information to each other to provide a complete picture of the situation. Developments in adjacent industries such as autonomous vehicles which require sensor fusion, and 360-degree vision are overlapping with the security industry as the silicon providers look to new markets for growth. AI based object filtering can reduce unwanted alarms that may be generated from basic video motion detection, making video surveillance more effective.

AI can add even more value when deployed in the cloud and make it possible to access various siloed data sources. APIs play an essential role in providing secure access to otherwise closed databases and allow analysis across multiple data sources. The benefits achieved from the use of AI range from lower storage costs to reduced unwanted alarm events and more useful intelligence. This means less time utilised reviewing and more time for action to reduce loss. The use of AI extends further into managing the operational uptime of video surveillance systems though predictive maintenance and management of product failures and the introduction of smart scheduling to improve engineer efficiency.

What does the future look like in terms of Video Surveillance?

I see a continuation of the progression towards the upgrade of older, low-quality cameras to higher resolution smart cameras which contain AI processing at the edge. Such cameras can be much more proactive in reducing and preventing crime. AIs have evolved to become extremely good at processing low grade images and add value without a hardware swap. However, it is inevitable that low-quality cameras will eventually fail as higher resolution will win-out and ultimately force legacy infrastructures to upgrade.

One recent example that comes to mind is the disorder at Wembley Stadium for the Euro 2020 final. Facial recognition technology could play a part in any future measures. A simple anonymised facial recognition check against tickets could have helped to avert the issue more quickly and safely. Providing information about the location of ticketless fans and enabling focused, strategic action by the security leadership. Upstream of the breach, predictive AI could be used to detect the presence of emerging threats and intent from sources gathered from outside the security industry technology ecosystem. e.g. social media and open source video feeds (traffic and satellite).

Wireless technology and 4G/5G networks open-up the potential to deploy video in more difficult or expensive to cable locations, whilst avoiding ingress and disruption of the corporate network. Drone and satellite video will play an increasing part in the rapid deployment of responsive solutions and counter terrorism activities.

The advent of cloud AI processing will lead to the increased use of video surveillance images for other purposes such as safety and processing monitoring. Video will become a data commodity which increases in value as it becomes more easily transferrable between applications. This opens a host of privacy and cyber security concerns for organisations. The strong component of video surveillance regulation contained within the proposed EU AI Act will result in greater accountability of technology providers and users and a shared responsibility throughout the supply chain. This should drive up the acceptance and use of AI based video surveillance products through transparency and trust in the way they are developed and used.

Recent high-profile breaches, such as the release of the footage of then Health Secretary, Matt Hancock, show that regulation is needed to serve to deter others who may seek to gain personally out of privacy breaches. Such action combined or aligned with a response from the industry, could serve to rebuild any lost trust in the video surveillance process arising from this case.


“We have used CSL via their various products for our alarm installs and we are now using their 4G routers for our remote area camera installs. Always a great product and Tech Support are available."

John Barclay (Saxen Fire & Security) – LinkedIn – 12th April

Other recent news