security analytics

Occupancy monitoring with thresholds that warn before a limit is reached

In short

Occupancy monitoring maintains a running count of people inside a defined space by combining directional entry and exit counts at every doorway. Thresholds raise a warning as a space approaches its limit and an alert when it is exceeded, so a supervisor can act before conditions become unsafe or uncomfortable.

Request a pilot Send a sample video

Conceptual illustration of stadium ingress and crowd management. AI-generated conceptual image, not a customer deployment.

Scope

What this analytic does, and what it does not

What it detects

Entries and exits at each monitored doorway
A derived running occupancy figure for the space
Threshold crossings against warning and limit values
Rate of change, so a rapid fill can be distinguished from a slow one

What it does not guarantee

An exact head count — occupancy is derived from counting and accumulates error over time
Correct totals when an unmonitored door is used
Any life-safety or evacuation function
Compliance with an occupancy regulation on its own

The problem

What customers are actually dealing with

These are the situations that lead teams to look at this analytic in the first place.

A venue only discovers a space is over capacity when it already feels dangerous.
Manual clicker counts at the door are inconsistent between staff and stop during the busiest moments.
Occupancy figures drift through the day and nobody knows whether the number on the dashboard is trustworthy.
Different rooms have different limits and there is no single view showing which are close to them.

How it works

Detection workflow

Every event carries the rule that produced it, so an operator can see why they were alerted.

1

Count at doorways

Directional crossing counts are produced at every entrance and exit to the space.
2

Aggregate

Entries and exits from all doorways are combined into a single running occupancy figure.
3

Evaluate thresholds

The running figure is tested against configured warning and limit values.
4

Stabilise

A short confirmation window prevents a brief threshold touch from raising an alert.
5

Deliver and reset

Alerts route to a supervisor, and scheduled resets re-baseline the count at known-empty times.

Configuration options

The set of doorways that bound the space, each with its own counting line
Warning and limit thresholds per space
Scheduled reset times when the space is known to be empty
Confirmation window before a threshold alert is raised
Whether staff-only doors are excluded from the occupancy calculation
Alert routing per space and per time of day

Alert quality

How irrelevant and duplicate alerts are reduced

No video analytic eliminates false alerts. These are the mechanisms that reduce them, and the residual rate is measured on your own cameras.

Scheduled resets at known-empty times are the single most important control on drift.
A confirmation window prevents one person stepping in and out of a doorway from toggling an alert.
Counting every bounding doorway is essential — an unmonitored door is the main cause of an untrustworthy figure.
Rate-of-change smoothing separates a genuine rapid fill from counting noise.
Occupancy remains a derived figure with accumulating error; the pilot measures drift over a realistic operating day.

Prerequisites

Camera requirements and environmental limits

Camera suitability decides more of the outcome than any software setting. These are assessed per camera before commitment.

Camera requirements

Every doorway bounding the space needs a camera view suitable for directional counting
An overhead or steeply angled view at the doorway is strongly preferred to reduce occlusion between people
Frame rate of 12–15 fps or higher at the doorway to capture fast entries
Consistent lighting at the threshold, including during evening operation
A view wide enough to see the full door width so nobody passes outside the counted line

Environmental limitations

Any unmonitored route into or out of the space invalidates the figure.
Groups entering abreast through a wide doorway are the main source of counting error.
Error accumulates through the day, which is why scheduled resets matter.
Prams, trolleys and carried children affect counting behaviour and should be considered in the acceptance test.

Delivery

Alerts, evidence and where they land

Event and evidence fields

Space name and configured limit
Current occupancy at trigger
Threshold crossed (warning or limit)
Contributing doorway counts
Timestamp with timezone
Time since last reset
Event identifier

VMS integration

Threshold alerts can be delivered as VMS events so an operator can view the doorway cameras immediately. The continuous occupancy series is usually more useful on an operational dashboard than inside the VMS.

See compatibility states →

Deployment options

Edge processing per building where doorway cameras are local
On-premise aggregation where a site has many bounded spaces
Hybrid deployment where per-space counts are aggregated into an estate dashboard

Compare architectures →

Measurement

Dashboard metrics and pilot acceptance criteria

Acceptance thresholds are agreed with you before the pilot starts. This site publishes no benchmark figures, because they do not transfer between sites.

Dashboard metrics

Current occupancy per space against its limit
Peak occupancy per space per day
Time spent above the warning threshold
Measured drift between scheduled resets
Entry and exit totals per doorway
Doorway camera availability

Measurable pilot criteria

Counting error per doorway measured against a manual count over defined busy and quiet periods
Occupancy drift measured between two known-empty resets across a full operating day
Threshold alert latency from the true crossing moment
Behaviour with a scripted group entry through a wide doorway
Completeness check that every bounding door is actually counted

Governance

Privacy and governance

Occupancy is an aggregate figure. No identity, face template or personal record is created.
Counting output can be retained as numbers only, with no imagery, where the use case allows.
Where occupancy data is used for staff areas, employee monitoring obligations may apply.

Frequently asked questions

How accurate is the occupancy figure?

It is a derived figure, not a head count. Accuracy depends on doorway camera views, how people enter (singly or abreast) and how often the count is reset. The pilot measures drift on your own doorways across a realistic operating day rather than quoting a general figure.

What happens if someone uses an unmonitored door?

The figure becomes wrong and stays wrong until the next reset. Identifying every bounding doorway during design is the most important step, and scheduled resets at known-empty times limit how long an error persists.

Can we use this for regulatory capacity compliance?

It can support an occupancy process, but it is a derived measurement and should not be presented as a compliance instrument on its own. Treat it as an operational early warning that prompts a human decision.

Keep reading

A controlled pilot establishes what this analytic actually does in your environment, against acceptance criteria we agree before it starts. Send a sample video first if you would rather see an assessment before committing to a pilot.