Skip to content
Camnitive
Book a demo

Use case · Safety

Fall & man-down detection

When someone goes down, the response clock starts: detection in seconds, not whenever a colleague walks by.

Book a demoTry it in the Rule Playground

The problem

Why it keeps happening.

A lone worker who falls in a stairwell, a remote pump station, or an aisle after hours may not be found for minutes or hours. Response time drives outcomes, and wearable panic devices only help if they are worn, charged, and reachable.

Camera coverage of these areas usually exists already, recording the event for later review rather than raising the alarm in the moment.

The rule

Written the way you'd say it.

In plain English
“Escalate to the control room immediately if a person falls and remains down for ten seconds.”
Compile it in the Rule Playground

How it runs

  1. 01

    Cover the risk areas you already film

    Stairwells, isolated aisles, remote assets, lone-worker areas, and unprotected edges connect via existing camera feeds. No wearables to distribute or maintain.

  2. 02

    Detect the fall, confirm the down state

    Detection tells a fall from bending or kneeling by confirming the person stays down past your threshold (seconds, tuned per area). Edge and opening zones can alert before a fall happens.

  3. 03

    Escalate immediately

    The control room or on-duty responder gets the alert with location and live context. Escalation tiers keep firing until acknowledged.

  4. 04

    Preserve the sealed record

    Every event is sealed and timestamped in FloFusion: response time and sequence are documented for the investigation and the insurer.

Measure

What's measured.

  • Falls and man-down events per area and shift
  • Time in edge zones, per area and shift
  • Time to first responder on scene

What it’s worth

  • Response time cut from discovery-by-chance to seconds
  • Lone-worker coverage without wearable programs
  • Edge exposure caught before it becomes a fall
  • Documented response for every event: a duty-of-care record

Estimate your ROI for your own operation, in about two minutes.

Comply

The evidence.

  • A clip and frame of the event, sealed with the device’s own key the moment it happens
  • The rule that fired, the zone and camera, and the time
  • The vision-language model’s verdict on the alert, with its reason
  • Who saw it, who acted, and when, in the audit log
FloFusion · alert review
An alert in FloFusion: the frame, the review choices, and the device seal and evidence check both verified

Common questions

How does it avoid false alarms from kneeling or sitting?

The person must stay down longer than a time you set before it alerts, and a vision-language model checks the frame first. Thresholds are tuned per area in the pilot.

Link to this question

Can it also warn before a fall?

Yes. Unprotected edges, floor openings and dock lips become zones, so anyone there, or there too long, raises an alert on the same cameras.

Link to this question

Does it work in low light?

It works wherever your cameras can see, so areas with night-capable or IR cameras are covered around the clock. Each camera is checked in the pilot.

Link to this question

More use cases

The use-case library

See it on your own cameras.

A pilot on the cameras you already own, with success criteria agreed up front.

Book a demoTry it in the Rule Playground