Combined sensing and video analytics for smarter security

Deploying outdoor security systems is a difficult task. In a complex and constantly changing environment, it is very challenging to find one single technology that can deliver it all – security, surveillance and categorisation.

At the Sicurezza exhibition, which took place from 22nd to 24th November, our team demonstrated a system that allied sensing and video analytics to create a bespoke and intelligent security system.

The demonstration consisted of our REDWALL SIP long range sensors, CCTV cameras and a NUC (Next Unit of Computing) running a Video Analytics software. The concept was to show how, by building on the strength of each technology, we could create specific event rules and responses.

For instance, our long-range volumetric sensors are very performant at detecting movement but can’t categorise the objects detected. The video analytics software running on the NUC or any other small form factor PC has the ability to categorise between animals, humans and vehicles. The system can be then configured to check what has been detected by the sensor and create different responses depending on the video analysis of the image. It could be set to ignore medium to large animals but trigger events for human and/or vehicle detection. Directional analysis can be added to the rules, so the system is alerted if the people or vehicles are moving in one direction and not the other.

Another scenario is the detection at night or in low light conditions. When it is dark, the camera field of view can be restricted, the quality of the video affected and therefore the capacity of the video analytics reduced. In such conditions, the system can be set to give priority to the sensor, if it triggers it raises an alarm, regardless of what the video analytics says. That way, the risk of missed alarms can be avoided.

While larger intrusion and surveillance systems rely on an Enterprise VMS or PSIM platform to manage complex event rules and logics, smaller installations could achieve a similar outcome by using cost-effective equipment such as a small form factor PC or NVR running video analytics software, standard CCTV cameras and OPTEX outdoor sensors.