LaneGuardVision.

🛣️👁️LaneGuard Vision

AI-Powered Highway Lane Enforcement System MVP

A real-time computer vision system built for Tech Sprint Hackathon to detect heavy vehicles illegally occupying fast lanes on National Highways.

---

🚨 LIVE DEMO (ACTION CAPTURE)

This project requires a Python backend to run AI models.

• Live Share Link: Click here to check if the session is active (Requires start_share.py running).
• Video Walkthrough: See demo.webp above.
• Backend Code: See the backend/ folder for the YOLOv8 + FastAPI logic.
• Frontend Code: See the frontend/index.html for the React-based dashboard.

---

🚀 Features

• Live Vehicle Detection: Uses YOLOv8 + OpenCV to detect Trucks, Buses, Cars, and People.
• Virtual Lane Enforcement: Monitors a defined “Fast Lane” zone and tracks vehicle duration.
• Smart Alerts: Triggers MQTT Alerts when a Heavy Vehicle stays in the fast lane for >5 seconds.
• Real-time Dashboard: A futuristic, dark-mode React UI (served via CDN) showing live video and violation stats.
• Dynamic Logging: Logs violations with timestamps and vehicle IDs.

🛠️ Tech Stack

• Backend: Python, FastAPI, Uvicorn
• AI/CV: YOLOv8 (Ultralytics), OpenCV
• Frontend: HTML5, React.js (CDN), TailwindCSS
• Data: MQTT (Paho), JSON

📦 Installation

1. Clone the repository:

   git clone https://github.com/your-username/LaneGuard.git
   cd LaneGuard
   

2. Install Dependencies:

   pip install -r backend/requirements.txt
   

   (Note: If strict versions fail, just install: ultralytics opencv-python fastapi uvicorn paho-mqtt)

3. Setup Models:
   • The app uses yolov8n.pt. If it’s missing, run:

     python download_model.py
     

🚥 How to Run

1. Start the Application: Double-click run_app.bat (Windows)

   OR run via terminal:

   python -m backend.main
   

2. Open Dashboard: Go to 👉 http://localhost:8000

🌐 Sharing with Evaluators (ngrok)

If you need to show the app to someone else (hackathon judges) without them installing anything:

1. Run run_app.bat to start the server.
2. Run python start_share.py in a separate terminal.
3. Copy the Public Link generated and paste it into the “Live Demo” section of this README.

⚙️ Configuration

📹 Video Source

• Webcam: To use your webcam, delete highway_sample.mp4 from the root folder.
• Video File: To use a video, place any MP4 file named highway_sample.mp4 in the root folder. (The app currently includes a sample video).

🚛 Detection Logic

Debug Mode (Current): The backend/detector.py is currently configured to detect Everything (Cars, Persons, Trucks) for demo purposes.

Strict Mode (Heavy Vehicles Only): Edit backend/detector.py:

# Change this line:
self.HV_CLASSES = [0, 2, 5, 7] 
# To this (Bus & Truck only):
self.HV_CLASSES = [5, 7]

📡 MQTT Alerts

Alerts are published to highway/enforcement/alerts on broker.hivemq.com (Public Broker).

📄 License

MIT License. Built for Hackathon purposes.