SafeLens AI
Real-time construction site safety system using computer vision.
85%+
mAP@0.5 PPE
25-30
FPS live video
<10%
No Hard Hat false negative
AI-powered site-safety monitoring for construction environments. SafeLens detects missing PPE, unsafe conditions, and high-risk behavior in real time from camera feeds using YOLOv8-based vision models.
Construction sites are high-risk environments where missing PPE and unsafe behavior often go unnoticed until an incident occurs. Manual monitoring is inconsistent and reactive rather than proactive.
I built a CCTV-first safety monitoring system powered by computer vision. The model detects hard hats, reflective vests, and unsafe acts in live footage and flags risky conditions before they escalate.
Pitched at GITEX Africa 2024. The project turns passive cameras into active safety infrastructure, giving site supervisors real-time visibility into hazards and compliance gaps.
Real-time PPE recognition for construction environments
Monitoring workflow designed for safety operations and incident prevention
Built with a strong operational focus on usable industrial deployment
