VLM output
A daytime CCTV view of a campus tree-lined grounds separated by a low metal railing from an adjacent sidewalk and street, with a red 'Virtual Fence' zone drawn over the grounds. In the first frame a person in dark clothing is detected climbing over the railing at the boundary; later frames show the walkway clear and then a man in a red shirt walking along the sidewalk away from the camera.
- Perimeter Intrusion DetectedYes
- Virtual Fence CrossingYes
- Fence Climbing BehaviorYes
- Persons Detected1-2
- Restricted Zone OccupiedNo
- Loitering DetectedNo
- Alert LevelHigh
- Lighting ConditionDaylight
- Crowd DensityLow
Customer challenges
Campuses are large, open, and busy, often with vulnerable populations; manual CCTV monitoring can't cover every camera, and incidents escalate before staff notice.
Customer benefits
24/7 proactive monitoring, faster response, safer students and staff, wide coverage of existing cameras, evidence capture, and reduced security workload.
Key features
Weapon Detection
Detects visible weapons in camera view and alerts campus security.
Incident Detection — Person Fall
Detects a person falling and raises an incident alert.
Fight / Aggression Detection
Detects fighting or aggressive physical behavior.
Unauthorized Access & Intrusion Detection
Detects entry into restricted or closed areas.
Loitering & Dwelling
Flags people lingering in an area longer than expected.
Fire & Smoke Detection
Detects fire and smoke early in monitored areas.
Crowd & Gathering Monitoring
Monitors crowd size and gatherings in campus spaces.
Before & after
| Before | After — with Linker AI Nexa |
|---|---|
| Before Incidents caught by chance on monitors or reported after they escalate. | After — with Linker AI Nexa Safety events flagged to security within seconds, with location, for human review. |