Light Alert

Light Alert is a mobile application designed to assist drivers by detecting traffic lights and notifying them when the light turns green. The app utilizes a custom-trained Core ML model to analyze a live camera feed in real time, providing sound and haptic feedback to alert the user.

---

🚀 Features

• Real-Time Traffic Light Detection: Detects traffic lights directly from the live camera feed.

• Green Light Notification: Alerts the user with sound and vibration when the light turns green.

• Simple User Interface: Minimalistic design for safe and distraction-free use.

• Welcome Screen with Instructions: Includes a guide to help users set up and use the app effectively.

---

🛠️ Technologies Used

Languages

• Swift:
  • Used for the entire app development, including SwiftUI for the user interface and Combine for state management.

Frameworks

• Core ML:

  • Powers the machine learning model for traffic light detection.

• Vision:

  • Processes the camera feed and passes frames to the Core ML model for inference.

• AVFoundation:

  • Manages the live camera feed for real-time detection.

• Core Haptics:

  • Provides tactile feedback when the light turns green.

• SwiftUI:

  • Implements a modern, declarative UI for a clean and simple user experience.

Custom Machine Learning Model

• The app uses a custom-trained Core ML model created with Create ML:
  • Trained on labeled images of traffic lights (red, yellow, green) and background scenes.
  • Optimized for real-time performance and high accuracy.

Others

• Combine:
  • Manages app state and data flow between the ViewModel and Views.

---

📱 How It Works

1. Set Up:

   • Mount your phone on your car's dashboard with a clear view of the road.

2. Detection:

   • The app processes the live camera feed and identifies traffic lights in real time.

3. Notification:

   • When the light turns green, the app provides:
     • Sound Notification: A "ding" sound.
     • Haptic Feedback: A vibration alert.

---

🚧 Future Enhancements

• Improved model accuracy with additional training data.
• Multi-language support.
• Advanced notifications for other traffic conditions.

---

🏗️ Project Structure

TrafficLightDetector/
├── Models/
│   ├── TrafficLightClassifier.mlmodel  # Custom Core ML model
├── ViewModels/
│   ├── CameraViewModel.swift           # Manages camera and detection logic
├── Views/
│   ├── WelcomeView.swift               # Welcome screen with instructions
│   ├── ContentView.swift               # Live detection view
│   ├── CameraView.swift                # Displays the live camera feed
├── Resources/
│   ├── AppIcon/                        # App icon assets
│   ├── ding.mp3                        # Notification sound
├── TrafficLightDetectorApp.swift       # App entry point
├── README.md                           # Project description (you’re here!)

---

📦 Installation

1. Clone the repository:

   git clone https://github.com/theabecaster/LightAlert.git

2. Open the project in Xcode:

   cd traffic-light-detector
   open TrafficLightDetector.xcodeproj

3. Build and run the app on a physical iOS device.

---

📝 License

This project is licensed under the MIT License. See the LICENSE file for more details.

---

👨‍💻 Author

• Abraham Gonzalez
  GitHub | Email

---

🌟 Acknowledgments

• Apple’s Core ML and Create ML for making machine learning integration seamless.
• Combine and SwiftUI for simplifying state management and UI development.
• The Xcode IDE for providing a streamlined development experience.