PPE, fire and smoke detection on 45 cameras at a furniture factoryAt a furniture and mattress factory, 45 existing cameras were connected to AI: hard-hat and vest (PPE) compliance plus fire and smoke are detected automatically, 24/7, across the site.2025-09-30
At a furniture and mattress factory full of foam, fabric and wood, both workplace safety and fire risk were high, yet inspection relied on a few people's eyes and periodic rounds. Hype Vision connected the plant's 45 existing cameras to a single platform over RTSP and built a system that detects hard-hat/vest (PPE) compliance, fire and smoke from those same cameras 24/7 — all fully on-premise.
The problem: high fire load + safety across a large area
In furniture and mattress production, foam, fabric, glue and wood dust create a high fire load; a small spark or smouldering can grow fast. At the same time, hard-hat and vest use by dozens of staff across a large production and storage area could not be monitored continuously.
The plant had 45 existing CCTV cameras but used them only for recording; no one could watch 45 screens at once.
PPE (hard-hat and vest) compliance on 45 cameras
People in each camera are detected and their hard-hat and hi-vis vest use is checked per person. Compliant staff are marked green, those without a helmet or vest red; a violation raises an alert with a timestamped evidence frame.
The dashboard shows live compliance by camera and area (e.g. helmet compliance, vest compliance) and the number of people and forklifts detected.
- Per-person hard-hat and vest checking
- Live PPE compliance by camera/area
- Timestamped evidence-frame alert on violation
- Count of people and forklifts detected
Fire and smoke detection across the site
The same cameras are used for fire and smoke detection. The model looks for flame and smoke signatures in the image and, by detecting them early, raises an instant alert via dashboard and WhatsApp — so in high-fire-load areas smoke is noticed before it grows.
Fire/smoke detection complements heat detectors: it catches smoke at its source, within the camera's field of view.
- Early flame and smoke detection from the image
- Instant alert via dashboard and WhatsApp
- Prioritisation of high-fire-load areas
- Visual evidence complementing existing detectors
45 cameras, one platform, on-premise
All 45 cameras are brought into one platform and each feeds both the PPE and the fire/smoke module at once. All analysis runs on a server inside the plant; video never leaves the site. People are not identified, tracking is anonymous and faces can be blurred.
Outcome
All 45 cameras are now not just recording but an active safety and fire-protection layer. PPE violations and fire/smoke risk are tracked on one screen with timestamped evidence, and response can start within seconds.
- 45 cameras on one on-premise platform
- Hard-hat/vest compliance logged 24/7
- Early fire and smoke warning
- Safety teams work from the same data
Related solutions
Frequently asked questions
Do 45 cameras need new hardware?
The cameras are existing and connect over RTSP/ONVIF. An edge server is placed inside the plant for analysis, sized during discovery by camera count and modules. One camera feeds both the PPE and the fire/smoke module.
Does fire/smoke detection replace detectors?
No, it complements them. The camera catches smoke and flame early within its field of view and is used alongside existing fire-detection systems.
Does PPE detection work at night or in low light?
It works as far as the camera can provide an image. In low light, lighting or a suitable camera is recommended; suitability is assessed in the free pre-assessment.
Does video leave the plant?
No. The deployment is on-premise; analysis runs on a server inside the plant and only event and metric data reach the dashboard.
