VisionAI: Real-Time PPE Detection and Behavior Monitoring

AI Summer of Code Project by <p> Club

VisionAI is an AI-powered surveillance system built to enhance workplace safety in industrial and high-risk environments. It uses real-time computer vision to detect Personal Protective Equipment (PPE) violations and monitor unsafe behaviors through live video feeds.

💡 What It Does

• Real-Time PPE Detection
  Identifies missing safety gear (helmets, vests, gloves) using YOLOv8.

• Instant Alerts
  Sends real-time notifications and plays voice warnings through camera speakers.

• Incident Logging
  Saves cropped violation images with timestamps and metadata for future reference.

• Unsafe Behavior Monitoring
  Detects abnormal actions such as running, falling, or limping through gait analysis.

• Access Control & Attendance
  Uses facial recognition to track attendance and restrict unauthorized access.

• Zone-Based Safety Enforcement
  Detects and logs restricted area violations.

⚙️ Tech Stack

• Computer Vision: YOLOv8, OpenCV
• Backend: FastAPI, SQLite (with future PostgreSQL upgrade)
• Frontend: React.js, Tailwind CSS
• Alert System: pyttsx3 for TTS, optional buzzer integration
• Storage: Local + Google Drive (for cropped violations)

📌 Why It Matters

Traditional CCTV systems are passive and rely heavily on manual review. VisionAI transforms this by actively enforcing safety rules, reducing response time, and ensuring accountability — making workplaces smarter and safer.

---

Stay tuned for updates as we continue to build a deployable, scalable, and intelligent safety compliance system.