PPE Detection
Automatically audit helmet, vest, glove and goggle use.
Whether a helmet was on is always obvious after the accident. The point is to have that information before it happens — on every shift and every person.
The model locates each person per frame, classifies the equipment worn and reports what is missing together with a confidence score. Auditing no longer depends on one inspector’s round; it becomes a rule that runs every second, everywhere a camera sees. Violations are counted anonymously and used as training and trend data, not as a disciplinary tool.
How it works
Person and equipment separated
The model finds the person, then looks for helmet, vest, gloves and goggles as separate classes.
Equipment assigned to the person
Detected equipment is matched to that person’s body box, preventing mix-ups in crowds.
Zone-specific requirements
Each area carries its own PPE rule: vest in the warehouse, goggles in the welding bay.
Where, when and on which equipment do violations concentrate?
The value of PPE data is not in individual violations but in their distribution. Once you see which gate people enter without a helmet, which shift compliance drops on and which item is consistently skipped, the response turns into a training plan, a layout change or a procurement decision — which is what makes it permanent.
| Helmet | 94% |
| Vest | 91% |
| Gloves | 78% |
| Safety goggles | 64% |
| Steel-toe footwear | 88% |
Glove and goggle compliance correlates directly with equipment comfort and accessibility; where locker positions were changed, the rate rose noticeably within the first month.
| W1 | 212 |
| W2 | 188 |
| W3 | 171 |
| W4 | 150 |
| W5 | 141 |
| W6 | 132 |
| W7 | 128 |
| W8 | 125 |
The first two weeks of the pilot are measurement, the rest is intervention; the decline holds because alerts reach the shift supervisor instantly.
Use cases
Line entrance
Turnstile alert and record on helmet-less entry.
Vest-mandatory area
Vest checks in corridors with forklift traffic.
Goggles and mask
Additional equipment rule in the chemical filling area.
Site entrance
The same standard applied to subcontractor staff.
Technical requirements
| Camera | Existing IP/CCTV; 1080p, person height ≥ 80 px in frame |
| Placement | Entrances, over-line and corridor views; steep top-down angles not recommended |
| Processing | On-premise server or cloud; one GPU server handles 8–16 cameras |
| Integration | Web panel, e-mail/SMS, WhatsApp, IoT alarm relay, REST API |
| Data | No facial recognition or biometric records; event clips retained for a set period |
Frequently asked questions
How does PPE Detection work?
Whether a helmet was on is always obvious after the accident. The point is to have that information before it happens — on every shift and every person.
Are our existing cameras suitable for this solution?
Required camera specification: Existing IP/CCTV; 1080p, person height ≥ 80 px in frame. Before installation we assess your camera inventory free of charge and report which points are suitable and, if needed, which camera should be added.
How accurate is it in the field?
Measured accuracy: 90%. Glove and goggle compliance correlates directly with equipment comfort and accessibility; where locker positions were changed, the rate rose noticeably within the first month. Thresholds are calibrated on your own site during the pilot.
Which systems does it integrate with?
Web panel, e-mail/SMS, WhatsApp, IoT alarm relay, REST API. Events are published over REST API and MQTT, so it connects to your existing MES, ERP, VMS and alarm infrastructure.
Is it compliant with KVKK (Turkish data-protection law)?
Footage is processed on an edge server inside your facility; raw video is not sent to the cloud. Facial recognition is not used; analytics outputs are anonymous and event clips are deleted automatically when the defined retention period ends.