Monitoring Station Activity with AI Enabled Smart Surveillance NVRs

Around the world, rail services are busier than ever before and catering for more passengers than they were originally intended for. This is creating a big problem for the security staff trying to safely manage the high passenger numbers at stations whilst stopping trains being boarded without a ticket and identifying suspicious or restricted individuals. Train stations regularly face potential security breaches and the security teams need to know of any unusual activity as it is happening.

Approached by one of Florida’s metro stations relying on an outdated security platform, operators are required to be watching multiple cameras via a control room to monitor any fraudulent or dangerous behaviour. This has proved hard work not only monitoring real-time activity but also for handing over the evidence to the local Police. As a result, they look towards the advances in AI NVR technology  in smart surveillance to help improve this.

“Footage that would previously have taken me hours to find, record, and share — now takes me a few minutes. The AI NVR provided by Things Embedded is very convenient and provides meaningful insights to our security camera footage for passing on to the relevant authorities.” Metro Station Security Guard

Monitoring activity through multiple screens in a control room is now a thing of the past. Security guards can be on the floor allowing them to respond faster to incidents with alerts being sent to their mobile devices of any programmed activity.

This level of video intelligence empowers security guards to not only make decisions based on real-time video data but historical backups are sent to the cloud. These insightful recordings are flagged for fast analysis of meaningful every day activities within the station allowing for improvements to be made in passenger safety. Here are a few security features supporting security guards through AI technology with the NVIDIA Jetson platform:

Virtual Fencing

Ability to monitor every individual passing through a defined area or through a virtual barrier. For example if someone were to cross through a ticket barrier in the wrong direction then the software will alert the user.

Facial Recognition

Powerful enough to recognise a large pool of traffic at a fast pace, the AI capabilities are set to detect matches from a managed watchlist of people of interest.

Classified Object Detection

Recognized objects will be flagged allowing operators to verify and respond faster. For example, operation of a skate board in high traffic area inside the building.

Behavioural Analysis

Trained to spot any potential dangerous behaviour, an alert would be sent if any suspicious activity is recorded.

No Face Mask Detection

To help prevent community transmission, this feature determines whether a person is not wearing a face mask.

Occupancy Counting

By placing cameras at exits and entrances, a running total of occupancy can be monitored in real time.