Use the intelligence already inside modern cameras
Many current IP cameras include onboard analytics. Where the camera model and metadata are supported, Luxriot can receive camera-side analytics metadata and display it over video, use it as the basis of Events & Actions rules, and make supported analytics events available for investigation in playback.
A site may already have cameras capable of useful people, vehicle, line-crossing or other analytics. A VMS migration can potentially bring those events into one operating environment instead of replacing cameras simply to obtain a new recorder feature.
What can analytics be used to detect?
Luxriot EVA uses neural-network object detection and classification. The generic detector can identify people, vehicles and animals, allowing rules to focus on relevant targets rather than basic pixel movement.
Create detection areas and virtual lines so events can be generated when an object enters, leaves, crosses or moves in a defined direction. This is useful for perimeter, entrance and restricted-area monitoring.
Counters can be linked to analytics rules to count people, vehicles or other supported objects. Counts can be used for live operational information and reporting.
Analytics can identify presence or dwell-type conditions where supported, helping turn long periods of uneventful video into a smaller number of events for an operator to review.
Luxriot's analytics ecosystem can support appeared/disappeared, abandoned or removed-object style rules where the selected analytics engine or camera-side metadata provides them. Capability should be confirmed against the proposed camera and analytics source.
Depending on the selected Luxriot analytics engine or integrated service, additional analysis can include PPE, plates, vehicle attributes, face-related searching and other specialist functions. Licensing and hardware requirements vary.
Camera-side Edge AI or server-side analytics?
The camera performs the analysis close to the image source. This can reduce the amount of central processing required and allows Videcom to make use of analytics already available in suitable cameras. Luxriot can consume supported edge metadata for overlays, events and search.
Luxriot EVA provides native neural-network analytics within the EVO ecosystem. EVA Basic includes object detection and familiar zones, lines, counters, metadata search and event rules, while more advanced analytics options are available where the application requires them.
Events & Actions turns a detection into a response
Detection is only useful if something happens next. Luxriot's Events & Action engine can use analytics events as triggers and combine them with schedules and other system information. A rule can be designed to alert an operator, display a camera or layout, create a log entry, send a notification, move a PTZ camera to a preset or activate an available output or integrated workflow.
Camera Edge AI, Luxriot EVA or an external service identifies an event.
Rules focus on the relevant object, zone, line, count or behaviour.
Luxriot Events & Actions applies the required operational response.
Operators search the recorded video using the associated analytics metadata and event history.
Bring camera-side analytics from different systems into one workflow
A mixed camera estate does not automatically mean one common analytics feature set. Different manufacturers expose different events and metadata. Luxriot's value is that supported camera-side metadata can be brought into the VMS alongside Luxriot analytics and other system events, giving operators a more consistent place to view, search and respond.
For supported integrations, Luxriot can also accept external analytics metadata in JSON format for video overlay, event triggering and search. This provides a route for specialist third-party analytics without forcing operators to work in a completely separate video application.
i-PRO Active Guard: a current example of deep Edge AI integration
Luxriot EVO 1.33 includes support for i-PRO Active Guard 3.0. Active Guard works with suitable i-PRO cameras with built-in AI and can provide detailed people and vehicle searching, attributes and watchlist-driven events inside the Luxriot operating workflow. This illustrates how edge intelligence can be combined with the VMS rather than treated as a stand-alone camera feature.
Useful applications for education and multi-site estates
Use object classification and defined zones to distinguish a relevant person or vehicle event from general scene movement, subject to correct camera positioning and configuration.
Apply analytics rules to selected entrances, compounds, car parks or restricted areas during defined times and feed those events into the operator workflow.
Use supported counters to provide information on movement through entrances, gates or other defined routes.
Search analytics events and metadata rather than reviewing long periods of continuous footage camera by camera.
Design matters: analytics performance depends on camera position, scene, lighting, object size, calibration, selected engine and manufacturer support. Videcom can test the proposed analytics against the real scene before treating it as an operational alarm source.
Tell Videcom what you want to detect, the camera makes already installed and how operators should respond. We can identify whether camera-side analytics, Luxriot EVA or an integrated analytics service is the best route.
Discuss Luxriot analytics