Disaster detection from aerial imagery with convolutional neural network

Siti Nor Khuzaimah Binti Amit, Yoshimitsu Aoki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

In recent years, analysis of remote sensing imagery is imperatives in the domain of environmental and climate monitoring primarily for the application of detecting and managing a natural disaster. Satellite imagery or aerial imagery is beneficial because it can widely capture the condition of the surface ground and provides a massive amount of information in a piece of satellite imagery. Since obtaining satellite imagery or aerial imagery is getting more ease in recent years, landslide detection and flood detection is highly in demand. In this paper, we propose automatic natural disaster detection particularly for landslide and flood detection by implementing convolutional neural network (CNN) in extracting the feature of disaster more effectively. CNN is robust to shadow, able to obtain the characteristic of disaster adequately and most importantly able to overcome misdetection or misjudgment by operators, which will affect the effectiveness of disaster relief. The neural network consists of 2 phases: Training phase and testing phase. We created training data patches of pre-disaster and post-disaster by clipping and resizing aerial imagery obtained from Google Earth Aerial Imagery. We are currently focusing on two countries which are Japan and Thailand. Training dataset for both landslide and flood consist of 50000 patches. All patches are trained in CNN to extract region where changes occurred or known as disaster region occurred without delay. We obtained accuracy of our system in around 80%-90% of both disaster detections. Based on the promising results, the proposed method may assist in our understanding of the role of deep learning in disaster detection.

Original languageEnglish
Title of host publicationProceedings - International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages239-245
Number of pages7
Volume2017-January
ISBN (Electronic)9781538607169
DOIs
Publication statusPublished - 2017 Dec 19
Event6th International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017 - Surabaya, Indonesia
Duration: 2017 Sep 262017 Sep 27

Other

Other6th International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017
CountryIndonesia
CitySurabaya
Period17/9/2617/9/27

Fingerprint

Disasters
Antennas
Neural networks
Satellite imagery
Landslides
Remote sensing
Earth (planet)
Monitoring
Testing

Keywords

  • aerial imagery
  • change detection
  • disaster detection
  • flood convolutional neural network
  • landslide

ASJC Scopus subject areas

  • Artificial Intelligence
  • Signal Processing

Cite this

Amit, S. N. K. B., & Aoki, Y. (2017). Disaster detection from aerial imagery with convolutional neural network. In Proceedings - International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017 (Vol. 2017-January, pp. 239-245). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/KCIC.2017.8228593

Disaster detection from aerial imagery with convolutional neural network. / Amit, Siti Nor Khuzaimah Binti; Aoki, Yoshimitsu.

Proceedings - International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017. Vol. 2017-January Institute of Electrical and Electronics Engineers Inc., 2017. p. 239-245.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Amit, SNKB & Aoki, Y 2017, Disaster detection from aerial imagery with convolutional neural network. in Proceedings - International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017. vol. 2017-January, Institute of Electrical and Electronics Engineers Inc., pp. 239-245, 6th International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017, Surabaya, Indonesia, 17/9/26. https://doi.org/10.1109/KCIC.2017.8228593
Amit SNKB, Aoki Y. Disaster detection from aerial imagery with convolutional neural network. In Proceedings - International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017. Vol. 2017-January. Institute of Electrical and Electronics Engineers Inc. 2017. p. 239-245 https://doi.org/10.1109/KCIC.2017.8228593
Amit, Siti Nor Khuzaimah Binti ; Aoki, Yoshimitsu. / Disaster detection from aerial imagery with convolutional neural network. Proceedings - International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017. Vol. 2017-January Institute of Electrical and Electronics Engineers Inc., 2017. pp. 239-245
@inproceedings{725d5157ce8c4f2d96f5936b500aa39a,
title = "Disaster detection from aerial imagery with convolutional neural network",
abstract = "In recent years, analysis of remote sensing imagery is imperatives in the domain of environmental and climate monitoring primarily for the application of detecting and managing a natural disaster. Satellite imagery or aerial imagery is beneficial because it can widely capture the condition of the surface ground and provides a massive amount of information in a piece of satellite imagery. Since obtaining satellite imagery or aerial imagery is getting more ease in recent years, landslide detection and flood detection is highly in demand. In this paper, we propose automatic natural disaster detection particularly for landslide and flood detection by implementing convolutional neural network (CNN) in extracting the feature of disaster more effectively. CNN is robust to shadow, able to obtain the characteristic of disaster adequately and most importantly able to overcome misdetection or misjudgment by operators, which will affect the effectiveness of disaster relief. The neural network consists of 2 phases: Training phase and testing phase. We created training data patches of pre-disaster and post-disaster by clipping and resizing aerial imagery obtained from Google Earth Aerial Imagery. We are currently focusing on two countries which are Japan and Thailand. Training dataset for both landslide and flood consist of 50000 patches. All patches are trained in CNN to extract region where changes occurred or known as disaster region occurred without delay. We obtained accuracy of our system in around 80{\%}-90{\%} of both disaster detections. Based on the promising results, the proposed method may assist in our understanding of the role of deep learning in disaster detection.",
keywords = "aerial imagery, change detection, disaster detection, flood convolutional neural network, landslide",
author = "Amit, {Siti Nor Khuzaimah Binti} and Yoshimitsu Aoki",
year = "2017",
month = "12",
day = "19",
doi = "10.1109/KCIC.2017.8228593",
language = "English",
volume = "2017-January",
pages = "239--245",
booktitle = "Proceedings - International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Disaster detection from aerial imagery with convolutional neural network

AU - Amit, Siti Nor Khuzaimah Binti

AU - Aoki, Yoshimitsu

PY - 2017/12/19

Y1 - 2017/12/19

N2 - In recent years, analysis of remote sensing imagery is imperatives in the domain of environmental and climate monitoring primarily for the application of detecting and managing a natural disaster. Satellite imagery or aerial imagery is beneficial because it can widely capture the condition of the surface ground and provides a massive amount of information in a piece of satellite imagery. Since obtaining satellite imagery or aerial imagery is getting more ease in recent years, landslide detection and flood detection is highly in demand. In this paper, we propose automatic natural disaster detection particularly for landslide and flood detection by implementing convolutional neural network (CNN) in extracting the feature of disaster more effectively. CNN is robust to shadow, able to obtain the characteristic of disaster adequately and most importantly able to overcome misdetection or misjudgment by operators, which will affect the effectiveness of disaster relief. The neural network consists of 2 phases: Training phase and testing phase. We created training data patches of pre-disaster and post-disaster by clipping and resizing aerial imagery obtained from Google Earth Aerial Imagery. We are currently focusing on two countries which are Japan and Thailand. Training dataset for both landslide and flood consist of 50000 patches. All patches are trained in CNN to extract region where changes occurred or known as disaster region occurred without delay. We obtained accuracy of our system in around 80%-90% of both disaster detections. Based on the promising results, the proposed method may assist in our understanding of the role of deep learning in disaster detection.

AB - In recent years, analysis of remote sensing imagery is imperatives in the domain of environmental and climate monitoring primarily for the application of detecting and managing a natural disaster. Satellite imagery or aerial imagery is beneficial because it can widely capture the condition of the surface ground and provides a massive amount of information in a piece of satellite imagery. Since obtaining satellite imagery or aerial imagery is getting more ease in recent years, landslide detection and flood detection is highly in demand. In this paper, we propose automatic natural disaster detection particularly for landslide and flood detection by implementing convolutional neural network (CNN) in extracting the feature of disaster more effectively. CNN is robust to shadow, able to obtain the characteristic of disaster adequately and most importantly able to overcome misdetection or misjudgment by operators, which will affect the effectiveness of disaster relief. The neural network consists of 2 phases: Training phase and testing phase. We created training data patches of pre-disaster and post-disaster by clipping and resizing aerial imagery obtained from Google Earth Aerial Imagery. We are currently focusing on two countries which are Japan and Thailand. Training dataset for both landslide and flood consist of 50000 patches. All patches are trained in CNN to extract region where changes occurred or known as disaster region occurred without delay. We obtained accuracy of our system in around 80%-90% of both disaster detections. Based on the promising results, the proposed method may assist in our understanding of the role of deep learning in disaster detection.

KW - aerial imagery

KW - change detection

KW - disaster detection

KW - flood convolutional neural network

KW - landslide

UR - http://www.scopus.com/inward/record.url?scp=85046417494&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046417494&partnerID=8YFLogxK

U2 - 10.1109/KCIC.2017.8228593

DO - 10.1109/KCIC.2017.8228593

M3 - Conference contribution

AN - SCOPUS:85046417494

VL - 2017-January

SP - 239

EP - 245

BT - Proceedings - International Electronics Symposium on Knowledge Creation and Intelligent Computing, IES-KCIC 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -