Applications / Security
Loitering & Reconnaissance Detection
Detects people, vehicles and craft that linger near fences, entrances and high-value assets beyond a normal pass-through time, and correlates repeat approaches across days.
See it in action
Dwell where nobody should be dwelling
Video slot — awaiting upload.
What it detects
Most hostile activity is preceded by an unremarkable period of standing around. Fogsphere measures dwell where dwell is not expected, and connects the repeats.
Fence-line loitering
Identifies people and vehicles remaining at or near the boundary beyond the configured pass-through time, per fence segment.
Entrance and forecourt loitering
Flags dwell at entrances, forecourts, back-of-house doors and service yards, including during trading hours where the zone is not a waiting area.
Abnormal dwell near sensitive assets
Surfaces time spent near cabinets, tool stores, high-value stock, control rooms and equipment enclosures that no task in the area requires.
Repeat-approach correlation
Connects the same person or vehicle appearing at the same location across separate days, so a pattern is visible rather than a series of unrelated alerts.
Vehicle loitering and under-vehicle activity
Detects vehicles standing at the boundary or in the car park, and activity beneath or around parked vehicles.
Drone loitering and small-craft standoff
Identifies drones holding position over the site and small craft holding off a platform, jetty or intake.
Concealment and suspicious behaviour
Flags concealment gestures, repeated re-entry and movement patterns inconsistent with the normal use of the space.
Trespass in restricted public space
Detects dwell inside restricted-zone boundaries in stations, depots, plant rooms and other publicly adjacent areas.
Out-of-hours presence
Surfaces people and vehicles present in a zone during hours the site is configured as closed, with dwell measured from first appearance.
Objects left after a dwell event
Links a loitering event to an object left behind at the same location, so the two are reviewed as one sequence.
Where it applies
Loitering detection appears in nine of the twelve Q3 papers, most explicitly in retail, oil & gas forecourt and offshore, smart-city, logistics and power-generation.
How it works
Edge inference on the cameras you already have
Dwell measurement runs on existing cameras at the edge. Because it is a time-based test on an already-tracked object, it adds little cost on top of detection that is running anyway.
Zones, thresholds and schedules you configure
Set the dwell threshold, the zone shape and the active hours per location. A forecourt at 14:00 and the same forecourt at 02:00 can carry entirely different definitions of loitering.
Every event carries its own evidence
Each detection is logged with a timestamp, the camera and zone it came from, and a short clip. The record can be reviewed, exported or attached to an investigation without trawling through hours of footage.
Feeds the systems already in place
Events surface in Fogsphere Manager and route outward to the VMS, access-control platform, control-room wallboard, ticketing queue or messaging channel operators already watch.
Risk & control objective
The control objective is early awareness, not confrontation. Between a site being observed and a site being entered there is usually a gap — someone standing at the fence, a vehicle parked on the access road, the same face on the forecourt three nights running. That gap is the only window in which anything can be decided calmly.
The measurement is deliberately simple: how long has this object been in a place where objects do not normally stay? What makes it useful is the threshold being local. A delivery driver waiting at a gate is not loitering; the same duration at a substation fence at midnight is.
The repeat-correlation half matters more than the single event. One person lingering is noise. The same vehicle at the same corner on four separate evenings is a report worth reading, and it is the record most sites have no way to produce.
Related solutions
See who is standing at your fence, and how often
We will show you dwell measurement and repeat-approach correlation on your own boundary and entrance cameras.