Overview

EODT4Crises is a comprehensive web-based platform for automated road detection and mapping from satellite imagery. Built in partnership with Helyx and funded by ESA, this tool combines state-of-the-art machine learning with an intuitive web interface to enable rapid road network extraction from multiple satellite data sources. The platform features an interactive Leaflet.js-based frontend with a Python backend, supporting real-time analysis and visualization of road infrastructure for applications in crisis response, and infrastructure monitoring.

Key Features

Interactive Web Interface

  • Leaflet.js-based Map: Interactive web mapping interface with smooth pan, zoom, and layer management
  • Real-time Road Detection: On-demand road extraction using the SAM_road model directly in the browser
  • OpenStreetMap Integration: Overlay and compare detected roads with existing OSM road networks
  • Multi-layer Visualization: Toggle between different data layers and analysis results

Multi-Source Satellite Data Integration

  • Google Earth Engine (GEE): Access to Landsat, Sentinel, and other satellite archives
  • Maxar: Access to sub-meter resolution satellite data for detailed analysis
  • Local File Upload: Support for user-provided satellite imagery and GeoTIFF files

Advanced Road Detection

  • SAM_road Model: Leverages the Segment Anything Model adapted specifically for road detection
  • Automated Vectorization: Converts detected road masks into clean vector geometries (GeoJSON, Shapefile)

Technology Stack

Frontend

  • Leaflet.js: Interactive web mapping library for smooth map interactions
  • HTML/CSS/JavaScript: Modern web technologies for responsive user interface

Backend

  • Python: Core processing engine for machine learning and geospatial operations
  • SAM_road: Specialized road detection model based on Segment Anything Model
  • Flask/FastAPI: Web framework for API endpoints and data processing

Applications & Use Cases

Crisis Response & Emergency Management

  • Rapid Assessment: Quickly assess road network damage after natural disasters
  • Evacuation Planning: Identify accessible routes for emergency response
  • Humanitarian Aid: Support logistics planning for aid delivery

Project Partnership

This project represents a successful collaboration between academic research and industry expertise:

  • Helyx: Parallel development for mapping of power-network infrastructure.
  • European Space Agency (ESA): Funding and supporting the development of this Earth observation tool

The partnership leverages ESA’s commitment to developing practical applications of satellite Earth observation data, combining cutting-edge AI research with real-world deployment challenges to create a tool that can make a meaningful impact in crisis response and infrastructure monitoring.