Detecting natural disasters, damage, and incidents in the wild

TL;DR

This paper introduces the Incidents Dataset, a large collection of annotated social media images, and demonstrates its use in detecting natural disaster incidents in real-world images, enhancing situational awareness.

Contribution

The work provides the first large-scale, annotated image dataset for incident detection and develops baseline models to identify disaster-related images in social media.

Findings

The Incidents Dataset contains 446,684 images across 43 incident types.

Baseline models can effectively detect incident images in social media data.

Filtering experiments show the dataset's utility in real-world disaster scene analysis.

Abstract

Responding to natural disasters, such as earthquakes, floods, and wildfires, is a laborious task performed by on-the-ground emergency responders and analysts. Social media has emerged as a low-latency data source to quickly understand disaster situations. While most studies on social media are limited to text, images offer more information for understanding disaster and incident scenes. However, no large-scale image datasets for incident detection exists. In this work, we present the Incidents Dataset, which contains 446,684 images annotated by humans that cover 43 incidents across a variety of scenes. We employ a baseline classification model that mitigates false-positive errors and we perform image filtering experiments on millions of social media images from Flickr and Twitter. Through these experiments, we show how the Incidents Dataset can be used to detect images with incidents in the wild. Code, data, and models are available online at http://incidentsdataset.csail.mit.edu.