EYEFIRE — Factory AI Safety System

Factory floors are high-risk environments, yet manual safety inspections can't watch everything simultaneously. EYEFIRE was built to give manufacturers a persistent, AI-powered safety officer that never blinks.

The system ingests RTSP camera feeds from existing factory CCTV infrastructure — no new cameras required. YOLOv8 models run inference in real time to detect Personal Protective Equipment (PPE): helmets, safety vests, gloves, and eye protection. Workers missing required PPE in designated zones trigger immediate alerts to floor supervisors via dashboard and mobile notification.

Behavioral safety analysis catches violations that PPE detection misses: workers entering restricted machinery zones, dangerous proximity to moving equipment, and ergonomically unsafe postures during manual handling. A separate model handles fire and smoke detection — response time from ignition to alert is under 1 second.

Golden Sea deployed EYEFIRE on NVIDIA Jetson edge devices co-located with the camera infrastructure. Edge deployment eliminates cloud upload latency and keeps video data on-premises — critical for clients with data sovereignty requirements. The React-based dashboard gives EHS (Environment, Health & Safety) managers a live plant-wide view with incident history, compliance trend reports, and shift-by-shift analytics.