UM-Deluge-Capstone / Machine-Learning-for-Damage-Assessments

A hands-on tutorial for object detection for urban planning professionals.

Machine Learning for Damage Assessments

A group of graduate student researchers from the University of Michigan collected and published this ML tutorial to aid damage assessment processes. This repository contains information and data for other researchers to replicate on their own. Using object detection algorithms, this tutorial will demonstrate one machine learning method for damage detection using YOLOv5.

• For information on how to design a street level damage assessment machine learning model, navigate to the "Damage Detection Tutorials" folder.

• For unannotated, raw images related to street level disaster damage, navigate to the "Images" folder.

Folder Contents

Damage Detection Tutorials
  Google Collaboratory Notebook
  Street-Level Damage Dataset
  Youtube Series

Images
  Hurricane
  Earthquake
  Tornado

For more information about the impetus of this project, please check out our Website (https://rise-above-the-deluge-um.webflow.io/)