The Impact of Artificial Intelligence to Video Surveillance

AI in the Industry of Surveillance

Traditionally, the video surveillance industry started off with many cameras looking at all sorts of things, but no-one looking at the camera footage. An overwhelming amount of feeds resulted in a re-active way of securing threats without the gathering of any insights.

The implementation of video analytics ensured a change of scenery. Hard-coded algorithms are looking at pixel changes and specific patterns, putting out simple alerts if any anomaly is detected. Next to pixel changes, the algorithms are also trained to do object classifications, based on specific geometric perimeters (f.ex a car has this shape, looks like that, …). Thus, to retrieve that kind of information, the camera needs very specific requirements and manual configuration (camera positioning, focal length, lightning, …). However, those results may cause some limitations.

First of all, the system might be triggered by too many similar perimeters, resulting in an overload and useless alerts. Moreover, this kind of video analytics needs to be applied selectively to certain cameras and depend on certain environments. Not being able to implement it where you want, it doesn’t make your whole installation a lot smarter. Thirdly, you might experience some technology incapability when trying to connect different technologies for cameras, software, etc.

Therefore, the video surveillance needed a new approach. Thanks to high performance IP cameras which develop high quality images, applications that have a strong compute power and GPU’s that support deep learning, Vaion have been able to create powerful security solutions by embedding AI into video analytics.

What is needed for AI in Video Surveillance?

There are a few conditions for really being able to leverage AI in video surveillance:

  • Being able to identify automatically objects, events and anomalies.
  • Software that isn’t too expensive. One needs to be able to scale the number of cameras per server in order to gather valuable insights, and thus the industry needs scalability in the deployment model (sometimes one will want to use public cloud, sometimes hybrid cloud or maybe he’ll want enterprise deployment).
  • Analysis in real time. The collection of images, analysis, detection, notification and alerts need to happen simultaneously, in order for the operator to work more efficiently.
  • Accurate and timely searches for when the operator can’t analyze in real time.
  • Presence maps for occupancy in rooms or energy efficiency.

These kinds of features combined, will make sure that you can offer your customers the most innovative intelligent security options in the marketplace. Vaion’s portfolio is built on them and their solutions fueled by machine learning will mean a great leap forward for video analytics as we know it today.

 

Interested to learn more about Vaion and meet them for a discussion?

Subscribe for one of the introduction webinars!

SUBSCRIBE HERE

 

Watch our videos below to learn more about Vaion’s core features

Vaion social distancing compliance Vaion vcore feature: Smart Presence Vaion vcore feature: Smart Search Vaion vcore feature: Spotlight

Interested?

Contact us at iot-analytics@techdata.be and arrange a personal meeting

Please follow and like us:

Geef een reactie

Het e-mailadres wordt niet gepubliceerd. Vereiste velden zijn gemarkeerd met *

Enjoy this blog? Please spread the word :)