Disaster Detection from SAR Images with Different Off-Nadir Angles Using Unsupervised Image Translation

Jian Song, Bruno Adriano, Naoto Yokoya

Research output: Contribution to journalConference articlepeer-review

Abstract

Synthetic aperture radar (SAR) images observed at different off-nadir angles have different intensities, and change detection methods using difference images do not work well. This problem hinders emergency response when there is no archive data with a consistent off-nadir angle as emergency SAR observation. In this paper, we investigate unsupervised image translation methods based on generative adversarial networks and autoencoders to detect flood and landslide areas using SAR images observed at different off-nadir angles. Comprehensive experiments of disaster detection using ALOS-2 PALSAR-2 images for three floods and two landslides show that the developed methods can significantly improve the accuracy of disaster detection using pre- and post-disaster images observed at different off-nadir angles.

Original languageEnglish
Pages (from-to)14-20
Number of pages7
JournalCEUR Workshop Proceedings
Volume3207
Publication statusPublished - 2022
Event2nd Workshop on Complex Data Challenges in Earth Observation, CDCEO 2022 - Vienna, Austria
Duration: 2022 Jul 25 → …

Keywords

  • Autoencoders
  • disaster detection
  • generative adversarial networks
  • image translation
  • off-nadir angle
  • synthetic aperture radar

Fingerprint

Dive into the research topics of 'Disaster Detection from SAR Images with Different Off-Nadir Angles Using Unsupervised Image Translation'. Together they form a unique fingerprint.

Cite this