RescU

Assessing Rescue Factors in Densely Built Urban Areas

Project Overview

This project explores the emergency rescue capabilities in densely populated urban environments through case studies in Mong Kok (Hong Kong) and Dhaka (Bangladesh). These areas are known for their complex layouts, narrow streets, and high building density.

The goal is to develop a "rescue factor"—a composite metric derived from geospatial analysis, image processing, and simulation—that quantifies the accessibility of different streets for emergency response. The resulting insights can support urban planning and emergency preparedness.

---

1. Data Collection and Extraction

• Satellite Imagery

  • Sentinel Hub
  • Google Earth Engine

• Street-View Imagery

  • Mapillary

2. Image Processing and Analysis

• Building Layout Identification

  • QGIS
  • GDAL

• Escape Route Detection

  • Darknet (YOLOv4)
  • TensorFlow

3. Proximity and Response Simulation

• Fire Station Proximity

  • GeoPandas

• Rescue Time Simulation

  • OpenRouteService

4. Rescue Factor Calculation

• Metric Development

  • NumPy
  • Pandas

• Statistical Analysis

  • Assess relationships with building height, street width, and population density.

5. Visualization and Reporting

• Interactive Mapping

  • Folium

• Dashboard Development

  • Plotly Dash
  • Streamlit

---

Dataset and Tools

Datasets

• Satellite imagery (Sentinel Hub, Google Earth Engine)
• Street-level imagery (Mapillary)
• OpenStreetMap data for road networks and fire station locations
• Optional: Real-time traffic data from OSM-compatible APIs

Tools

• Image Processing: OpenCV, YOLOv5/Faster R-CNN
• Geospatial Analysis: QGIS, GDAL, GeoPandas, Shapely
• Routing Simulation: OpenRouteService
• Visualization: Folium, Streamlit, Dash

---

📄 License

To be determined.