Skip to content

README.md

Latest commit

 

History

History
228 lines (158 loc) · 5.46 KB

README.md

File metadata and controls

228 lines (158 loc) · 5.46 KB

🛣️ RoadGuard — AI-Powered Road Distress Detection

Python AI SAM License

RoadGuard is an AI-driven road infrastructure monitoring system designed to automatically detect and segment road distresses (such as potholes) using computer vision and deep learning. It supports geo-tagged detection, mask-based visualization, and is designed for web & mobile deployment.

📑 Table of Contents

📌 About the Project

Road infrastructure degradation is a major challenge for transportation authorities. Manual inspection is slow, expensive, and error-prone.

RoadGuard leverages AI & ML to:

  • Detect road surface damage
  • Segment potholes using masks
  • Store GPS-based location data
  • Enable scalable road condition monitoring

❗ Problem Statement

Traditional road inspection methods:

  • Require manual surveys
  • Are not scalable
  • Lack real-time insights
  • Miss early-stage deterioration

💡 Solution Overview

RoadGuard provides an automated AI solution that:

  1. Takes road images or video frames
  2. Detects distress regions
  3. Segments potholes using Segment Anything Model (SAM)
  4. Saves geo-coordinates for mapping and maintenance planning

✨ Key Features

  • 🚧 AI-based road distress detection
  • 🎯 Mask-level segmentation of potholes
  • 📍 Latitude & longitude tagging
  • 🖥️ Web & mobile-friendly architecture
  • 📊 Visual overlays for severity estimation

🧰 Tech Stack

Layer Technology
Language Python
Computer Vision OpenCV
Deep Learning Segment Anything Model (SAM)
Frontend TypeScript / React
Backend Python (ML inference)
Mapping GPS coordinates

🏗️ System Architecture

Input Image / Video
        ↓
AI Detection Model
        ↓
SAM Segmentation
        ↓
Mask Overlay + Geo-Tagging
        ↓
Web / Mobile Visualization

🧠 Model & Methodology

  • Detection Stage Identifies road distress regions from images.

  • Segmentation Stage Uses Segment Anything Model (SAM) to extract precise pothole boundaries.

  • Post-Processing

    • Mask visualization
    • Area estimation
    • Location tagging

⚙️ Installation

git clone https://github.com/Quantamaster/RoadGuard.git
cd RoadGuard

Create a virtual environment:

python -m venv venv
source venv/bin/activate   # Linux/Mac
venv\Scripts\activate      # Windows

Install dependencies:

pip install -r requirements.txt

▶️ Usage

python main.py
  • Provide road images or video frames
  • Model outputs segmented pothole masks
  • Location data is stored for mapping

📊 Results & Outputs

  • ✔ Accurate pothole segmentation
  • ✔ Mask-based visual overlays
  • ✔ Geo-referenced road distress points
  • ✔ Ready for integration with GIS dashboards

🌍 Applications

  • Smart city infrastructure monitoring
  • Highway maintenance planning
  • Road safety analysis
  • Automated inspection systems
  • Government & municipal agencies

🚀 Future Improvements

  • Real-time video stream inference
  • Severity scoring & prioritization
  • GIS dashboard integration
  • Edge deployment (mobile devices)
  • Multi-class road defect detection

📊 Project Insights (Interactive HTML)

RoadGuard includes a rich, interactive HTML-based insights page that presents the project as a portfolio-grade, business-ready AI system.

🔹 What’s Inside the Insights Page

  • Executive-style project overview
  • Key performance metrics (FPS, confidence, cost reduction)
  • Feature breakdown with visual hierarchy
  • System architecture & technical highlights
  • Business impact and real-world applications
  • Future roadmap and scalability vision

🔹 View Insights

🔗 Interactive Insights Page:
https://quantamaster.github.io/Akhilesh-Y-Ragate.github.io/Projects/roadguard-insights.html

GHBanner

🚀 Run and Deploy your AI Studio App

This contains everything you need to run your app locally.

View your app in AI Studio: https://ai.studio/apps/drive/1tGdz-pSLCFx9TQlYB64oAhsblhRGXGrF

Run Locally

Prerequisites: Node.js

  1. Install dependencies: npm install
  2. Set the GEMINI_API_KEY in .env.local to your Gemini API key
  3. Run the app: npm run dev