PPE detection using your existing CCTV cameras

The problem

How it works

1. 1Video streams are read from existing IP cameras or the NVR over RTSP/ONVIF.
2. 2A deep-learning model detects each person and classifies the PPE items configured for that zone.
3. 3Zone, duration and schedule rules decide when a detection becomes an event (for example, "no hard hat for more than 5 seconds inside the press area").
4. 4Events are sent to the dashboard and notification channels with a short evidence clip or snapshot, and are stored for reporting.

What can be detected

• Missing hard hat / helmet
• Missing high-visibility vest
• Other PPE items such as safety glasses, gloves or masks — evaluated per site and camera view
• Repeated violations in the same zone or time window
• PPE compliance trends by zone, shift and day

Typical scenarios

• Entrances to production halls where PPE is mandatory
• Loading docks and warehouse aisles with vehicle traffic
• Construction or maintenance areas with temporary rules
• Turnstile or door control where entry is refused when required PPE is not detected

Working with existing cameras

• Standard 1080p IP cameras are usually sufficient when the person occupies a reasonable part of the frame.
• A slightly elevated, angled view works better than a straight top-down view for helmet and vest classification.
• Small items (glasses, gloves) need closer framing; suitability is checked on sample footage before the pilot.
• Most IP cameras and NVRs that provide RTSP or ONVIF streams can be connected, regardless of brand.
• Before a pilot, sample footage from each candidate camera is reviewed for resolution, angle, distance and lighting.
• Where a view is not suitable, the recommendation is usually to reposition the camera or add one — not to replace the whole system.

Integration

• Web dashboard and mobile notifications
• REST API / webhook for EHS, ERP or ticketing systems
• Relay or PLC signals for sirens, beacons or turnstiles

Privacy and security

• No face recognition is used in safety and operations analytics; events describe what happened, not who.
• On-premise processing keeps video inside your network.
• Evidence clips, retention periods and user access are configurable and should follow your GDPR/KVKK policies.
• Workplace monitoring should be introduced with clear employee information and, where applicable, consultation with employee representatives.

Limitations

• Detection quality depends on resolution, distance, lighting and occlusion; it is measured per camera during the pilot.
• The system identifies a PPE violation, not a person. It does not use face recognition.
• Rules must reflect site policy — the same item may be mandatory in one zone and optional in another.