Artificial Intelligence and Video Surveillance – Part 3
- Security Solutions
- Feb 11
- 3 min read
For this time together we are going to explore where AI lives and the benefits and challenges of each style. There are three main places where AI can be processed; at the edge or on the camera itself, at the camera server which is still onsite or at least on the customers network, and finally the cloud. To complicate this a little more than can be combinations of these.
THE EDGE
Edge analytics are when the cameras process the images on the camera itself. The main advantage is the manufacturer can build the camera with the feature set in mind and make sure the processor meets those requirements. Since the camera really doesn’t change it means the feature set included will continue to work the way it was designed versus a shared resource like a network video recorder. The disadvantage of this is there is no way to upgrade the camera features since the processing is all being done at the camera itself.
CAMERA SERVER or NETWORK VIDEO RECORDER (NVR)
As mentioned above the advantage to this kind of system is the ability to upgrade the analytics system wide and the servers can be designed with incredible processing power. Given systems have been designed this way for decades there are a lot of resources that can be deployed to supplement the native system. Major disadvantages are the cost to deploy cameras outside the original design and overall costs.
CLOUD
The cloud can mean so many things. We are going to discuss two ways the cloud has been used for AI and Analytics. The first way would be to supplement the existing recording methods. This is typically done by the camera or NVR being the first layer of analytics and sending information to the could to be processed by the larger supercomputers. This is very effective at adding features since we are not relying strictly on the onsite features the NVR or Cameras have. The disadvantage to this is connectivity, your systems are now relying on the cloud to process potentially critical alerts.
The second method is fully cloud. In this method the camera is buffering a bit of video and streaming this to the cloud. The positives are impressive as all cloud features are available on all the cameras all the time. The downsides to this include bandwidth requirements, costs, and latency.Â
AI is a powerful tool in surveillance and individual needs will dictate the type of system needed. There are no one size fits all situations. If you believe you have an opportunity to implement AI in a way that can help you or your company let me know and I’ll help think through the solution with you.
With more than three decades in the security industry, Jamie has built a career spanning hands-on technical work and high-level leadership. He began as a locksmith, later expanding into life safety and eventually into electronic and mechanical security solutions. This well-rounded background gives him a unique, practical perspective on keeping people and property safe.
Jamie has served as president of both the Washington State and national Electronic Security Association and was honored with the industry’s highest recognition—the Weinstock Person of the Year award. A respected voice in the field, Jamie is a frequent speaker at national conferences and continues to advocate for smarter, safer solutions in schools, businesses, and communities.