Logo image
Back
Post-Disaster Damage Detection using Aerial Footage: Visual Question Answering (VQA) Case Study
Conference proceeding

Post-Disaster Damage Detection using Aerial Footage: Visual Question Answering (VQA) Case Study

Rafael de Sa Lowande, Arash Mahyari and Hakki Erhan Sevil
2022 IEEE Applied Imagery Pattern Recognition Workshop (AIPR)
IEEE Applied Imagery Pattern Recognition (AIPR 2022) (DC, USA, 10/11/2022–10/13/2022)
04/10/2023
DOI: https://doi.org/10.1109/AIPR57179.2022.10092229
Web of Science ID: WOS:000991969300031

Metrics

154 Record Views
1 Times Cited - Web of Science

Abstract

Natural disasters are a major source of significant damage and costly repairs around the world. After a natural disaster occurs, there are usually an insurmountable amount of damage along with financial costs of repairing and aiding all the people involved. Besides that, the occurrence of natural disasters has increased significantly in the past decade. With that in mind, post-disaster damage detection is usually performed either in person or manually by human experts. Taking into consideration all the areas one has to closely look into, as well as the inaccessible terrain, debris, and unstable infrastructure, it is incredibly difficult for a surveyor to identify and annotate every single possible damage out there. Previous studies in damage detection from Unmanned Aerial Vehicles (UAVs) have lead to great outcomes. Yet, these algorithms do not collaborate with human experts. This paper develops a Visual Question Answering (VQA) technique for post-disaster damage detection on aerial footage. Our case study on the dataset collected after the hurricane Sally shows promising results.

Details

Logo image