Artificial Intelligence no longer means vast server rooms that power the likes of IBM Watson or AlphaGo. Since AI analytics takes place on devices throughout our physical world, streaming data to a server is not required for all AI actions. In light of this, it makes sense that the market is expected to grow at a compound annual growth rate of 19.27 % between 2019 and 2026, according to the ‘Edge AI Market’ report.
In network surveillance systems, AI at the edge promises a variety of benefits, but let’s examine Hanwha’ Techin’s top five that make the biggest difference to installers and end users.
- Accuracy and context
Artificial intelligence can improve the accuracy of triggers and reduce false alarms. By using deep learning, edge AI can accurately count people, measure occupancy, manage queues, and more. Additionally, edge devices will provide operators with contextual information, like whether someone is wearing glasses or carrying a bag, as well as the colour of the vehicle.
- Ownership costs decrease
Approximately 2,500 petabytes of data are generated daily by security cameras globally. As sending all this data back to the cloud for storage and analysis is costly, using AI at the edge will immediately reduce data transfer costs (since only critical data and metadata needs to be sent back), as well as the cost of maintaining expensive servers. This also leads to energy savings.
- Latency is reduced
Using analytics at the edge, triggers and alerts can be sent more quickly. It improves the speed of response when dealing with an event, such as trespassing, and the overall experience when accessing or entering a site.
Thomas J. Bittman, a distinguished VP Analyst at Gartner, explains, “As people need to interact with their digitally-assisted realities in real-time, waiting on a data centre miles (or many miles) away isn’t going to work. Latency matters. I’m here, right now, and I’m gone in seconds.”
- Greater scalability
Cameras with edge AI will allow a video installation to be both flexible and scalable because more cameras and devices can be added when the installation needs change (without having to invest in a large server with significant bandwidth right at the start). This is especially helpful to organisations that wish to deploy a project in stages over a period of time.
Because analytics are being done across different devices, if one device fails, another can take over. There is no single point of failure in the system. The AI can continue to operate even if a network or cloud service fails. It is still possible to act on triggers, grant access, etc., with local recordings being sent to the back end as soon as the connections have been restored, and the front-end can still gather insights.
Find out more about Hanwha here.