🪖 Helmet Detection System

An AI-powered helmet detection system using YOLOv8 for safety compliance monitoring with Arduino hardware integration.

🚀 Features

• AI-Powered Detection: YOLOv8-based helmet detection with high accuracy
• Real-time Processing: Live webcam detection with instant results
• Hardware Integration: Arduino control for motor, LED, and buzzer
• Safety Compliance: Automatic safety checks with hardware responses

📁 Project Structure

demo-helmate/
├── helmet-detector/
│   ├── data/
│   │   ├── data.yaml              # Dataset configuration
│   │   ├── train/                 # Training images (3,648)
│   │   ├── valid/                 # Validation images (372)
│   │   └── test/                  # Test images (149)
│   ├── models/
│   │   └── helmet.pt              # Trained YOLOv8 model (6.2MB)
│   ├── scripts/
│   │   └── detect.py              # Main detection script
│   ├── input/                     # Test images
│   └── helmet_control_fixed.ino   # Arduino code
├── yolov8n.pt                     # Base YOLOv8 model
├── requirements.txt               # Python dependencies
└── README.md                     # This file

🛠 Installation

Prerequisites

• Python 3.8+
• Arduino IDE
• Webcam (for live detection)

Setup

1. Install Python dependencies

   pip install -r requirements.txt

2. Verify installation

   python -c "import ultralytics, torch, cv2; print('✅ All dependencies installed')"

🎯 Usage

Quick Start

1. Test with a single image

   cd helmet-detector
   python scripts/detect.py input/helmate-on-1.webp

2. Start live webcam detection

   python scripts/detect.py --webcam

Detection Results

• ✅ PASS: Helmet detected → Motor ON, LED OFF, Buzzer OFF
• ❌ FAIL: No helmet → Motor OFF, LED ON, Buzzer alert
• ⚠️ CONFLICT: Conflicting detections → Treat as FAIL
• ❓ NO_DETECTION: No objects detected → Treat as FAIL

🔌 Arduino Setup

Hardware Components

• Arduino Uno (or compatible)
• Motor (via relay) - simulates bike engine
• Red LED (with 220Ω resistor) - warning indicator
• Buzzer - alert system
• USB Cable - for serial communication

Wiring Diagram

Arduino Uno:
├── Pin 8 → Relay → Motor (Engine control)
├── Pin 9 → 220Ω Resistor → Red LED (Warning light)
├── Pin 11 → Buzzer (Alert system)
└── USB → Computer (Serial Communication)

Connection Steps

1. Connect Hardware

   • Connect motor to Pin 8 via relay module
   • Connect red LED to Pin 9 with 220Ω resistor
   • Connect buzzer to Pin 11
   • Connect Arduino to computer via USB

2. Upload Arduino Code

   • Open Arduino IDE
   • Open helmet-detector/helmet_control_fixed.ino
   • Select correct board and port
   • Click Upload

3. Find COM Port

   • Open Device Manager (Windows)
   • Look for "Arduino" under "Ports (COM & LPT)"
   • Note the COM port number (e.g., COM5)

4. Update COM Port in Script

   • Open helmet-detector/scripts/detect.py
   • Find line with serial.Serial('COM5', 9600, timeout=1)
   • Change 'COM5' to your actual COM port

Arduino Code Overview

// Pin definitions
const int motorPin = 8;     // Motor control
const int ledPin = 9;       // Red LED
const int buzzerPin = 11;   // Buzzer

// Signal handling
if (data == '1') {
    // Helmet detected: Motor ON, LED OFF, Buzzer OFF
    digitalWrite(motorPin, LOW);   // Motor ON
    digitalWrite(ledPin, LOW);     // LED OFF
    digitalWrite(buzzerPin, LOW);  // Buzzer OFF
} else if (data == '0') {
    // No helmet: Motor OFF, LED ON, Buzzer ON
    digitalWrite(motorPin, HIGH);  // Motor OFF
    digitalWrite(ledPin, HIGH);    // LED ON
    // Buzzer alert sequence
}

📊 Model Performance

Dataset Statistics

• Training Images: 3,648
• Validation Images: 372
• Test Images: 149
• Classes: 2 (With Helmet, Without Helmet)

Test Results

• Helmet Detection: 97.7% confidence
• No Helmet Detection: 88.4% confidence
• Average Inference Time: ~200ms per image

🔧 Configuration

Model Parameters

• Confidence Threshold: 0.5
• Camera Resolution: 640x480
• Detection Classes: With Helmet, Without Helmet

Arduino Settings

• Baud Rate: 9600
• Timeout: 1 second
• Port: COM5 (configurable)

🐛 Troubleshooting

Common Issues

1. Camera not detected

   • Close other applications using camera
   • Try different camera indices (0, 1, 2)
   • Check camera drivers

2. Arduino connection failed

   • Check COM port in Device Manager
   • Update Arduino drivers
   • Try different COM ports (COM3, COM4, COM5, COM6)
   • Ensure Arduino code is uploaded

3. Low detection accuracy

   • Ensure good lighting
   • Position 2-3 feet from camera
   • Avoid shadows on face
   • Check image quality

4. Model loading error

   • Verify helmet.pt exists in models/ folder
   • Check file permissions
   • Reinstall ultralytics if needed

Arduino Troubleshooting

1. Motor not responding

   • Check relay connections
   • Verify Pin 8 connection
   • Test relay with multimeter

2. LED not lighting

   • Check 220Ω resistor connection
   • Verify Pin 9 connection
   • Test LED polarity

3. Buzzer not working

   • Check Pin 11 connection
   • Verify buzzer polarity
   • Test with simple tone() function

📝 Usage Examples

Image Detection

# Test helmet detection
python scripts/detect.py input/helmate-on-1.webp

# Test no helmet
python scripts/detect.py input/helmate-off-1.jpg

Live Detection

# Start webcam detection
python scripts/detect.py --webcam

Expected Output

============================================================
HELMET DETECTION TEST RESULTS
============================================================
Total detections found: 1
🟢 With Helmet detected (confidence: 0.98)
Result: ✅ Helmet worn correctly

FINAL TEST RESULT
🎯 ✅ HELMET WORN CORRECTLY
   Safety compliance: PASSED ✅
   Status: SAFE TO PROCEED

⚠️ Safety Notice

This system is designed for educational and demonstration purposes. Always follow local safety regulations and wear appropriate protective equipment.

📞 Support

For issues and questions:

1. Check the troubleshooting section
2. Verify all connections
3. Test components individually
4. Check Arduino serial monitor for debugging

---

Happy detecting! 🪖✨