Glossary
The language of industrial safety AI.
Every term used in AI-powered workplace safety monitoring, explained plainly.
- AI workplace safety monitoring
- Using computer-vision AI on camera feeds to detect workplace hazards in real time. It covers PPE breaches, vehicle-pedestrian conflicts, falls, fire and restricted zone violations, generating a live alert that enables a response before an incident occurs rather than recording it for review afterwards.
- Vehicle-pedestrian detection
- AI detection of dangerous proximity between moving vehicles (forklifts, HGVs, reach stackers, terminal tractors) and people on foot. This is the struck-by hazard that is the leading cause of UK workplace fatalities. Alert latency is measured in under one second from the moment the proximity threshold is crossed.
- PPE detection
- Automatically checking from camera footage whether workers are wearing the required personal protective equipment in the zones that require it. This covers hard hats including colour verification, hi-vis vests, gloves, safety glasses, ear defenders, harnesses and safety boots.
- Red zone / exclusion zone
- A defined exclusion or danger area around a hazard: a moving machine, a suspended load, a pressurised vessel, an active drill floor operation. AI red-zone monitoring flags anyone who enters it without authorisation or the required PPE, in real time.
- Moonpool
- An opening in the hull of a drillship or semi-submersible offshore platform through which the drill string, riser and subsea equipment are deployed. Moonpool monitoring detects people entering the high-risk area during operations using exclusion zones that activate automatically.
- On-premise (edge) processing
- Running the AI analysis locally, on a compute device installed within the customer network, rather than sending video to a cloud service. On-premise processing means raw video never leaves the site, detection latency is measured in milliseconds, and the system continues to function during internet outages.
- Alert fatigue
- The tendency for operational crews to ignore or disable a safety monitoring system that produces too many false alerts. Alert fatigue is the primary mechanism by which safety technology fails in practice. The sub-0.05% error rate maintained by SecureSafety keeps alert rates low enough for every alert to be taken seriously.
- Near-miss
- An unsafe event that did not cause injury on this occasion. Near-miss data is the most valuable leading indicator of workplace safety performance. AI monitoring generates near-miss data automatically and continuously.
- Discovery phase
- The scoped technical assessment that precedes every SecureSafety deployment. It covers a camera survey, hazard mapping, network and integration assessment, DPIA framework, and a fully costed delivery plan. Discovery is a fixed-scope, fixed-price engagement that produces a board-ready proposal before any hardware is committed.
- DPIA
- Data Protection Impact Assessment — a documented assessment required under GDPR/LGPD before beginning systematic monitoring of individuals in a workplace context. SecureSafety provides a DPIA framework template as part of the Discovery phase documentation.
See it in action
See the detectors running on a live deployment.
Book a demo and we'll show the technology at work — real hazards, real cameras, live.