Skip to content
Camzify
AI detection · Heatmap anomalies

Heatmap anomalies

See where people really go. Heatmap anomalies maps foot traffic across a site and flags patterns that deviate from the established baseline — unusual congestion, empty zones, or unexpected activity.

Book a demo All 22 detections
  • Foot traffic congestion above the normal pattern for a zone
  • Activity in a normally low-traffic area outside expected hours
  • A monitored area sitting unusually empty during typically…
Site floor plan overlaid with a colour-coded foot traffic heatmap highlighting an anomalous zone
Heatmap anomalies

This capability detects and alerts on:

  • • Foot traffic congestion above the normal pattern for a zone
  • • Activity in a normally low-traffic area outside expected hours
  • • A monitored area sitting unusually empty during typically busy hours
  • • A visual heatmap overlay for layout and staffing decisions
  • • Zone-to-zone comparison to spot where traffic is shifting on a site
  • • Anomaly history reviewable alongside other site events

Why heatmap anomalies matters

Most foot traffic on a site is unremarkable, people moving through as expected, at roughly the volume you'd expect for the time of day. The interesting moments are the exceptions: a corridor that's suddenly congested, a normally busy lobby that's gone quiet, a back area with activity at an hour when nobody should be there. Those exceptions are easy to miss without something actively watching for them.

A single fixed rule, "alert if more than N people are in this zone", doesn't capture what "unusual" actually means for a given area, because normal varies by zone, by hour, and by day of week. A number that's alarming in a back corridor at 2am is completely ordinary in a lobby at lunchtime.

Heatmap anomalies solves this by comparing current traffic against a baseline built specifically for each zone, rather than a single threshold applied everywhere. That's what lets it flag a real deviation instead of either missing it or flooding the queue with false alerts.

Diagram showing current zone traffic compared against a learned baseline pattern before an anomaly is flagged
Baseline vs deviation logic

How it works

Building the baseline

Confirmed subject counts from multi-object tracking are aggregated per zone over time to build a baseline traffic pattern for each area of the site.

Detecting a deviation

Current traffic is continuously compared against that baseline. A deviation beyond the configured threshold, in either direction, flags an anomaly with the affected zone and time window.

Where the data feeds

Heatmap data also feeds the platform's analytics and reporting module for trend review independent of any single anomaly.

Configuration

Zones are marked on the camera view or site map. Each zone supports:

  • • Zone boundaries marked on the camera view or site map
  • • Anomaly sensitivity threshold, configurable per zone
  • • Baseline learning period before anomaly flagging goes active
  • • Notification window per camera, e.g. notify in business hours only
  • • Per-camera instance licensing
Configuration panel showing a monitored zone drawn on a site map with an anomaly sensitivity threshold control
Zone & sensitivity configuration
Dashboard view listing flagged heatmap anomalies by zone, time window, and deviation severity
Anomaly review

Common scenarios

  • • A retail aisle showing unusual congestion outside a promotional period
  • • A back-of-house corridor with unexpected foot traffic after closing
  • • A lobby sitting unusually empty during a normally busy morning window
  • • A queue forming in an area not designed for queuing, flagged for layout review
  • • A traffic pattern shift near a restricted-adjacent zone worth a security look
  • • A seasonal deviation from baseline that operations teams review before adjusting staffing

In a patrol round

Heatmap anomalies run as continuous background analytics rather than a per-camera checklist item during a virtual patrol round, but a flagged anomaly at a patrolled site is visible alongside that site's patrol reports.

Industries using this

RetailRestaurantsEducationFinancial ServicesProperty Management

Related detections

Occupancy & Peak Hour TrendsMotion Detection

Use cases

Guard Tour VerificationIncident Investigation
FAQ

Frequently asked questions

The system builds a baseline foot traffic pattern for each monitored area over time. An anomaly is a deviation from that baseline, unusual congestion, an empty zone that's normally busy, or unexpected activity in a low-traffic area — flagged for review rather than a single fixed rule.

The baseline improves as more traffic data accumulates for each zone. Early results reflect a smaller sample; accuracy improves over the first few weeks of continuous monitoring as normal patterns become established.

Both. Security teams use anomalies to flag unusual activity outside expected patterns, while operations teams use the same heatmap data for layout, staffing, and queue-management decisions.

An anomaly is flagged for review in the notification queue with severity and acknowledgment status, the same as other alert types. Whether it prompts a security response, an operations look, or no action at all depends on the zone and context, which is why it's surfaced for a human decision rather than acted on automatically.

Each camera carries a notification window, so anomaly notifications can be limited to the hours that matter and a planned event or seasonal rush doesn't generate a flood of expected-but-flagged notifications. The detection itself keeps running; the window decides when it tells anyone.

The two are complementary. Heatmap anomalies flags deviations from a zone's established baseline, while occupancy and peak hour trends tracks live counts and historical patterns for staffing and planning decisions. Anomalies answer "does this look unusual"; trends answer "when is this normally busy".

Ready to patrol your site 24/7?

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

Book a DemoCalculate Your Savings

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.