Vicon’s enhanced line of Roughneck AI Cameras and its latest VMS release—Valerus 23.1—combine to deliver powerful AI-based object classification, tracking, recording, and forensic searching. This tight integration between the cameras and Valerus reinforces Vicon’s established position as an end-to-end solutions provider.

The Roughneck AI Cameras—with features like adaptive IR, Starlight low-light technology, and -40° operating temps—now include built-in AI-based analytics that leverage deep learning.

The evolution of deep learning in the video surveillance space is significant, as it allows a camera to classify object types so users can expedite investigations and create intelligent real-time alerts. Additionally, deep learning systems can be continuously trained and improved with better—and more—datasets. Many applications have shown that deep learning systems can “learn” to achieve near-human-level accuracy for certain tasks.

“Vicon’s AI-based analytics feature object classification, allowing them to intelligently identify people and vehicles, help users perform more meaningful searches in Valerus, and diminish those nuisance alarms triggered by irrelevant motion,” said Bob Germain, Director of Camera Products. “The ability to filter out non-relevant video allows for faster incident resolution and lets users respond to genuine threats quicker and with fewer resources,” continued Germain.

Taking it a step further, Valerus can perform object-based recording, meaning motion-triggered events will be recorded only when they involve a person and/or vehicle. Object-based recording, and the ability to filter out “environmental noise” that can trigger a motion alarm (e.g., shadows, swaying trees and bushes, etc.), significantly increase operational efficiencies.

Users can leverage the new AI filters in Valerus 23.1 in several ways. For example, with Museum Search, users can apply object classification filters and limit their results to just people and/or vehicles, as opposed to “motion in general.” Additionally, users can search for events based solely on the objects they specify. For example, with an intrusion analytic set up, users can be alerted only when people (as opposed to vehicles or other objects) enter the intrusion zone. This delivers more meaningful search results, and ultimately reduces the amount of non-relevant video users have to watch.

For customers who perform live monitoring, these new object classification analytics combine with Vicon’s standard suite of analytics to trigger real-time meaningful alarms. It’s more than just a motion detection alarm; it’s motion that’s triggered specifically by a person and/or vehicle. Customers will be notified—in real-time—about events that are important to them, so live monitoring becomes easier, and they can manage more cameras.

These edge-based analytics are performed on the camera, which ultimately saves customers money and allows them to be more efficient with their server resources. There’s no need for additional costly processing hardware, and sending less video over the network minimizes the network load and reduces storage requirements.

“Another huge customer benefit is that these cameras can perform continuous recording in low-res (e.g., 720p) and then switch to hi-res (e.g., 4K) when object-based motion is detected,” said Louis Rabenold, Director of VMS Products. “This way, recording is always available but doesn’t put unnecessary strain on the customer’s storage and bandwidth requirements,” added Rabenold.

Vicon’s AI solution is truly plug-and-play right out of the box. “Customers will love how they can connect their Roughneck AI camera to Valerus 23.1, and they’re ready to go,” said Rabenold. “There’s no setup required on the camera in order to send/receive metadata, or perform object-based recording and museum search.”