SJUK Masthead Black

Knowing right from wrong – analytics and the future of surveillance

November 3, 2021


Video surveillance, as we know it, is an outdated way of thinking about protecting an organisation’s people and property. Many are lulled into a false sense of security just because they’ve installed cameras around their premises. However, simply recording activity isn’t enough.

It’s time to move toward using surveillance material in a smarter and more targeted way, that will allow us to identify and act upon unusual or suspicious behaviour without first having to trawl through hours of footage.

Companies need to understand the key moments to watch, what they mean and how to take action during an incident. This is why security needs to be geared toward an integrated approach; whereby video cameras, access controls and sensors work together, underpinned by advanced analytics and cloud technologies, to collect data intelligently and provide the whole picture.

AI-powered analytics for intelligence and insight

The use of traditional video surveillance systems means security professionals are faced with the task of sifting through vast amounts of unstructured – and often irrelevant – data each day. Even if an incident occurs, it might be overlooked as a result of operators receiving the wrong data at the wrong time.

Intelligent analytics are key to making sense of this data and providing operators with the critical insight they need. Enhanced by machine learning, these analytics are able to comb through hours of footage to find the critical security moments and make sense of them. AI algorithms enable software to “learn” what’s considered normal for an organisation and what’s not. This enables raw video content to be converted into actionable intelligence and enhanced situational awareness – far more valuable to a security professional than an uninterrupted stream of mostly meaningless video data.

For example, AI-powered analytics can detect, identify, extract and catalogue video input data such as object type, appearance and similarity. The inclusion of behavioural detection factors like dwell times, line crossings, vehicles in no-cross zones, or the presence of people in sensitive areas adds further context. Also, directional sound analytics on gunshots, screams or breaking glass mean operators only get alerts on incidents that really matter.

Applying advanced analytics such as these across all an organisation’s security cameras will allow operators to enjoy more of a complete picture. It’s here, then, that we should look to the cloud.

Connecting cameras to the cloud

The cost and difficulty of integrating several disparate cameras into a single cohesive system can be prohibitive for many organisations. However, connecting them to a modern, cloud-based architecture can enable them to enjoy the greater coverage afforded by a joined-up system as well as the benefits of the latest analytics technology.

Importantly, especially in the wake of COVID, where many people still work remotely, connecting an organisation’s security cameras to the cloud means footage can be analysed and watched from anywhere – even from home. It also means businesses no longer need to depend on traditional on-premise storage solutions. Again, this is important in the post-pandemic period, with many offices currently left unoccupied. Instead, all video recordings, along with essential event markers like bookmarks and alarms, can be stored in the cloud. Organisations can, therefore, feel confident in their security posture, even when their premises are unattended.

Surveillance technology has developed in recent years. A cloud-based security system, augmented by AI-powered analytics, puts more complete data at the fingertips of an organisation’s security professionals – wherever they are – and, by only saving those events deemed to be “unusual” or out of the norm, can also enhance their storage.

By learning, adapting to and making sense of unstructured data, AI algorithms can get a whole picture of what normal behaviour looks like and what indicates risk. In this way, by identifying the most important incidents – often as they happen – organisations are able to take immediate action, protecting their people and properties with proactiveness and intelligence never previously possible. Missing important events due to being overwhelmed by data is now a thing of the past.

Sam Lancia

By Sam Lancia, Head of Video Engineering & Co-Founder of Ava Security

For more information, visit:

Read Next

Security Journal UK

Subscribe Now

£99.99 for each year
No payment items has been selected yet