Skip to content
Camzify
AI detection · Slip & fall detection

Slip & fall detection

A fall happens, help gets notified fast. Slip and fall detection flags falls in real time and routes an alert to the nearest guard — before an incident goes unnoticed and becomes an unresolved claim.

  • Rapid, uncontrolled posture changes consistent with a fall
  • A person remaining on the ground after a fall event
  • Falls in aisles, entrances, and walkways during business hours
Camera view of a retail aisle with a fallen person highlighted by a detection bounding box
Slip & fall detection

This capability detects and alerts on:

  • • Rapid, uncontrolled posture changes consistent with a fall
  • • A person remaining on the ground after a fall event
  • • Falls in aisles, entrances, and walkways during business hours
  • • Falls near stairs, ramps, or wet-floor areas where risk is elevated
  • • A fallen individual who doesn't get back up within a short window
  • • Immediate alert routing to the nearest available guard

Why slip and fall detection matters

Falls tend to happen when no one is looking directly at the spot where it occurs, an aisle with no other customers nearby, a hallway between patrol rounds, an entrance during a quiet stretch. The gap between the fall and someone noticing is exactly the time when a minor injury can become a serious one, and it's the hardest part of the incident to control with staffing alone.

The usual backstop is the person themselves calling for help, or a passerby happening to notice. Both are unpredictable, someone who is injured, disoriented, or elderly may not be able to call out, and low-traffic areas or off-hours periods can go long stretches without anyone walking through at all. A scheduled patrol round covers a location for a moment every so often, not continuously.

Continuous AI monitoring removes the dependency on a witness being present. The moment a tracked subject's posture matches a fall pattern, an alert reaches the nearest guard directly, cutting the time between the event and a response, and producing a timestamped record of exactly what happened, independent of whether anyone saw it live.

Diagram showing a tracked subject's posture change over time crossing the fall-detection threshold
Fall pattern detection

How it works

Tracking posture over time

Built on multi-object tracking, the model watches each tracked subject's posture over time rather than evaluating a single frame in isolation.

Confirming a fall pattern

A rapid downward posture change followed by the subject remaining low or motionless crosses the alert threshold. Every alert includes a clip of the moments before and after the fall, timestamp, and location, so the responding guard has full context before arriving.

Alert delivery

Alerts route through the platform's notification system with priority routing to the nearest assigned guard.

Configuration

Slip and fall detection is enabled per camera. Configurable options include:

  • • Operating hours during which the detection is active
  • • Sensitivity threshold for what counts as a fall pattern
  • • Priority routing rules to the nearest assigned guard
  • • Zone exclusions, e.g. seating or exercise areas where lying down is expected
  • • Per-camera instance licensing
Configuration panel showing sensitivity and priority routing settings for slip and fall detection
Fall detection configuration
Facility map showing camera coverage across aisles, entrances, and walkways for fall detection
Fall risk monitoring

Common scenarios

  • • A grocery aisle where a customer slips on a wet floor and doesn't get back up
  • • A hospital or care facility corridor where a patient falls out of a chair
  • • A stairwell landing where a fall could otherwise go unnoticed for a long stretch
  • • An apartment building lobby or hallway during a quiet period
  • • A restaurant dining area during off-peak hours with few staff on the floor
  • • A building entrance during icy or wet weather conditions

In a patrol round

Slip and fall detection runs continuously rather than only during scheduled checks, but an event during an active virtual patrol round is logged immediately as a non-compliance event in the patrol report.

FAQ

Frequently asked questions

The model looks at the speed and pattern of a person's posture change and whether they remain on the ground afterward, rather than a single frame. A brief crouch or bend doesn't match the pattern; a rapid, uncontrolled drop followed by staying down does.

An alert with a timestamped clip routes immediately to the nearest available guard through the notification system, so someone can respond in person or dispatch help, and the event is logged with a timestamp for incident records.

Yes. Every detected fall produces a timestamped clip that documents exactly what happened and when, which is useful both for getting help to the person quickly and for producing an accurate record if a claim is filed later.

Detection quality depends on the camera's own low-light performance rather than the fall-detection model itself. Any feed with enough visibility for a human reviewer to make out a person's posture is enough for the tracker to register a fall pattern.

It works with existing camera coverage of aisles, walkways, and entrances. Cameras with a clear view of the ground plane in the monitored area give the most reliable results, but no dedicated fall-detection hardware is required.

A wearable depends on the individual choosing to wear it, keeping it charged, and it functioning correctly at the moment of the fall. Camera-based detection covers anyone who passes through the monitored area without requiring them to opt in or wear anything.

Ready to patrol your site 24/7?

Book a 15-minute demo and see a live patrol run on your own cameras.

This site is being updated

We are rebuilding pages as you read this, so an image, a link or a section may look unfinished for a while. The product itself is unaffected. If something important is broken, tell us at the contact page and we will fix it.