Sentinel - Safe Vision with AI (SafeGuard)
SafeGuard uses AI-powered PPE detection and IoT-based worker authentication to improve industrial workplace safety.
- Video of it working
- Problem, approach and result explained
- Expert reviewed· 6 reviews
—
Real-time PPE detection using YOLOv8 and AI.
—
Secure worker authentication using fingerprint and NFC.
—
Improved workplace safety through monitoring and access control.
Walkthrough
Overview
The problem
this problem mainly affects workers and safety officers in industries like construction, manufacturing and mining. It matters because missing PPE can lead to serious workplace injuries.
The approach
we built SafeGuard, an AI-powered industrial safety monitoring system using YOLOv8 and OpenCV for real-time PPE detection. It processes live camera feeds through a local server, while ESP32, fingerprint and NFC authentication help identify workers and support access control.
The result
the result is a working prototype that combines AI-based PPE detection with IoT-based worker authentication. It can monitor safety compliance through a live camera feed and display detection results. It demonstrates how traditional CCTV monitoring can become more intelligent and proactive.
Reflection
this project taught us that building a real-world AI solution requires more than just coding. We learned the importance of dataset quality, detection accuracy, hardware integration, teamwork and continuous testing. We also understood that practical challenges help us improve our solution.
Submitted to Real-Time Factory Safety Vision
- Your idea in one line
Our project transforms existing CCTV cameras into intelligent, real-time(24/7) safety guardians that automatically detect PPE violations, fire, and smoke to prevent workplace without requiring new hardware.
- What will be working by the finale?
Our final demonstration should prove that , the system helps to protect every worker's life by taking immediate actions against the problem quickly and accurately.
- Tech stack and hardware you plan to use
Frontend - HTML, CSS , Javascript Language - Python Model - YOLOv8 Framework - Numpy, OpenCV Notification - Telegram Notifier, telegram bot Streaming - MJPEG over http, RTSP
- What have you built or tested so far?
AI Safety detection: worker and PPE (Helmet, Goggles, Vest, Fire and smoke) detection using YOLOv8 models Camera integration: RTSP stream and USB webcam input support , Live Dashboard. Incident Management: Safety Alerts, incident logs , and CSV/JSON report export.
- Pitch deck (PDF)
- Pitch deck (PDF)PDF
- What's working now?
1.We develop an application for our site. 2. Customize PPE for different zones. 3. Implemented text to speech notification. 4. Optimize our site to support multiple CCTV (cameras). 5.Implemented IoT multilevel authentication system (NFC, fingerprint,Face recognition).