Zones, direction and dwell time along the fences of industrial sites, solar farms and data centres.
VCA Module
AI-powered video analytics
An analytics module for NetStation, built on technology from the British company VCA Technology — one of the pioneers of video analytics. Deep learning recognises people, vehicles and objects, and event rules turn video into alarms and metadata.
VCA ModuleVCA Technology, integrated with NetStation- analytics functions
- 24
- channels with VCA analytics
- 1.5M+
- VCA Technology on the market since
- 2009
How it works
- 01Camera imageVCA Module analyses video from cameras connected to NetStation.
- 02AI analysisDeep learning detects, classifies and tracks people, vehicles and objects.
- 03Event and metadataA matched rule triggers an alarm in NetStation, and metadata feeds the recording search.
Person 97%
Car 94%
Bag · abandoned 2:14
CAMERA 04 · PARKING · 10/14/2026 9:42 PM
Deep learning instead of motion detection
Classic motion detection reacts to shadows, rain, animals and moving branches. Deep-learning trackers recognise what an object is — a person, a vehicle or an item — so alarms reach the operator only when they matter.
- Fewer false alarms
- Classification of people, vehicle types and objects
- Tracking even with partial occlusion
Zone: FenceAND
Person (DL)AND
Dwell > 30 sNOTSchedule: 7:00–17:00
Alarm in NetStation: intrusion outside working hours
Dwell > 30 sNOTSchedule: 7:00–17:00Logic rules tailored to your site
Individual rules can be combined into scenarios — zone, object type, dwell time and schedule together. Alarms then reflect real threats, not every movement in the frame.
- Combine rules with logical operators
- Custom alarms tailored to the site
- Events go straight to NetStation
Vehicle · cameras 01–08 · yesterday 6–10 PM
VehicleZone: EntranceDirection: inbound
18:04
18:31
19:12
19:47
20:05
20:58
21:20
21:42
Metadata that cuts search time from hours to minutes
For every frame, the analytics record a description of detected objects: type, position, direction and time. After an incident there is no need to scrub through footage — just search for a person, vehicle or object in selected cameras and time range.
- Full metadata stream (JSON)
- Search by object type, zone and direction
- Faster preparation of evidence
All analytics functions
Event rules
When the system should react — where, in which direction and for how long.
- ZonesDraw areas of any shape on the image where the system should react.
PresenceAlarm when an object is inside a zone — e.g. a person in a restricted area.
EnterEvent when an object enters a zone from outside, across its edge.
ExitEvent when an object leaves a zone — e.g. a vehicle leaving the premises.
AppearAn object appears inside a zone without crossing its edge — e.g. emerging from behind an obstacle.
DisappearAn object disappears inside a zone — e.g. hiding or getting into a vehicle.
DirectionReact only to movement in a chosen direction — e.g. wrong-way driving.- DwellAlarm when an object stays in a zone longer than a set time — loitering detection.
StoppingDetects objects that have stopped — e.g. a vehicle in a no-stopping area.
SpeedFilters objects moving faster or slower than a set threshold.
Abandoned objectDetects items left unattended or removed from the scene.
Camera tamperAlarm when a camera is covered, moved or blinded, or loses video.
CountingCounts people and vehicles by type, even in crowded scenes.
Counting linesCounts line crossings in both directions — building entries and exits.
Logic rulesCombine rules into complex scenarios, e.g. zone + object type + schedule.
Tracking and AI
Who or what is in the frame and what it is doing — recognised by deep learning.
Object trackingA robust tracker follows objects across the frame, even with partial occlusion.- People (DL)Deep-learning people tracking — no alarms from shadows, animals or vegetation.
Objects and vehicles (DL)Classifies vehicle and object types, e.g. car, motorcycle, bicycle, bag.
Skeleton (DL)Human pose analysis — the basis for fall and behaviour detection.
Hand and objectDetects hand-to-object contact — e.g. taking goods off a shelf.
Aggressive behaviourRecognises violent movements that may indicate a fight or assault.
Fall detectionAlarm when a person falls — e.g. in a hospital, care home or warehouse.
Configuration and data
More accurate analysis and faster search in recordings.
3D calibrationPerspective calibration lets you filter objects by real size and speed.
MetadataPer-frame object descriptions (JSON) — the basis for fast search in recordings.
Where VCA Module fits
Customer counting, queue analysis and detection of interaction with goods.
Vehicle direction, stopping and speed; abandoned items at terminals.
Fall and aggressive-behaviour detection in hospitals and care homes.
Public-space monitoring with fewer false alarms.
Find events in recordings using metadata instead of manual review.
