🛡️ PPE Detection & Smart Access Control System

This project is an AI-powered safety system that performs real-time PPE (Personal Protective Equipment) detection, face recognition, and Arduino-controlled access using a stepper motor, LEDs, and a webcam. It's designed to ensure only authorized individuals wearing the required safety gear (helmet, mask, and vest) can enter a secure area — such as a lab, construction site, or industrial facility.

---

🎯 Key Features

• ✅ Real-time person and PPE detection using a YOLOv8 model
• 🎥 Live webcam feed with bounding boxes and confidence scores
• 🧠 Face recognition for logging personnel identity
• 🎙️ Voice feedback using pyttsx3
• 🔄 Arduino integration with:
  • Green/Red LED indicators
  • Stepper motor-controlled door/gate
• 📝 Excel log export of recognized persons and timestamps
• 🚫 Automatic denial if multiple people are detected simultaneously

---

🧰 Hardware Requirements

• Arduino Uno (or compatible)
• 4-Wire Stepper Motor
• Green & Red LEDs
• Webcam
• USB cable (for Arduino)
• Breadboard & jumper wires
• Compatible PPE (Helmet, Mask, Vest) for detection model

---

🛠️ Software & Libraries

• Python 3.8+
• Ultralytics YOLOv8
• OpenCV (opencv-python)
• cvzone
• pyfirmata
• pyttsx3
• pandas, openpyxl
• Custom YOLO model: ppe.pt
• Face Recognition Utility: SimpleFacerec

---

🗂️ Project Structure

PPE_Detection_Project/
├── PPEDetection-arduino.py         # Main Python script
├── Images/                         # Folder for known face images
├── ppe.pt                          # Trained YOLOv8 model for PPE
├── recognition_log.xlsx            # Auto-generated log file
├── simple_facerec.py               # Face recognition utility
└── README.md

---

▶️ Running the Project

1. Install dependencies

pip install ultralytics opencv-python cvzone pyfirmata pyttsx3 pandas openpyxl

2. Upload StandardFirmata to Arduino

Open Arduino IDE → Tools → Board: Arduino Uno → Upload StandardFirmata from Examples.

3. Run the main script

python PPEDetection-arduino.py

The system will prompt the first user to enter. Only one person is allowed at a time.

---

⚙️ How It Works

1. Webcam captures live feed.
2. YOLO model detects:
   • Hardhat
   • Mask
   • Safety Vest
3. Face is recognized using SimpleFacerec.
4. If only one person is present and all PPE are worn:
   • 🔊 "Access granted"
   • ✅ Green LED lights up
   • 🔁 Stepper motor rotates to open the door
5. Otherwise:
   • ❌ Red LED is lit
   • 🔊 "Only one person allowed" or PPE warning

---

📦 Output Example

• Real-time display with annotated detections
• Audio feedback
• Logged entries:

| Name       | Time                |
|------------|---------------------|
| John Doe   | 2025-06-01 09:31:15 |

---

📌 Customization

• Add authorized face images in Images/ folder.
• Modify detection classes or confidence threshold in code.
• Replace ppe.pt with an updated or more specialized model.

---

📜 License

This project is licensed under the MIT License.

---

🙋 Author

Developed by Rayan Ali Tlais
GitHub: ryn2004t(https://github.com/ryn2004t)
Email: [tlsryn2@gmail.com]

---

❤️ Contributions

Pull requests, improvements, and suggestions are welcome!