Challenge of Roof Damage Housings Detection from Satellite Images by Applying Deep Learning Methodology: -A Case Study of Ibaraki City at 2018 Osaka Earthquake-

Munenari Inoguchi, Seiichi Kara, Kazuya Shirai, Atsushi Imai

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

In Japan, we were suffered by many kinds of disasters. Once disaster occurs, we have to develop the common operational picture including damage situation in order to realize effective disaster response. However, it should take much time-cost to gather the damage situation. Against this issue, we decided to detect blue sheets object put on the damaged roof in recovery phase of disaster response. In this research, we tried to detect damage situation from satellite images by utilizing deep learning methodology. Especially, we adopt VGG-16 model developed by Oxford university, which gained fourth prize of ILSVRC in 2014. We prepared training data and applied it to actual affected area by 2018 Osaka earthquake as a case study. Finally, we confirmed that our trained AI detected blue sheet object with about 95% accuracy ratio.

Original languageEnglish
Title of host publicationITC-CSCC 2020 - 35th International Technical Conference on Circuits/Systems, Computers and Communications
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages172-176
Number of pages5
ISBN (Electronic)9784885523281
StatePublished - 2020/07
Event35th International Technical Conference on Circuits/Systems, Computers and Communications, ITC-CSCC 2020 - Nagoya, Japan
Duration: 2020/07/032020/07/06

Publication series

NameITC-CSCC 2020 - 35th International Technical Conference on Circuits/Systems, Computers and Communications

Conference

Conference35th International Technical Conference on Circuits/Systems, Computers and Communications, ITC-CSCC 2020
Country/TerritoryJapan
CityNagoya
Period2020/07/032020/07/06

Keywords

  • common operational picture
  • deep learning
  • disaster response
  • roof damage
  • satellite image

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Hardware and Architecture
  • Information Systems and Management
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Challenge of Roof Damage Housings Detection from Satellite Images by Applying Deep Learning Methodology: -A Case Study of Ibaraki City at 2018 Osaka Earthquake-'. Together they form a unique fingerprint.

Cite this