Post-Disaster Building Database Updating Using Automated Deep Learning: An Integration of Pre-Disaster OpenStreetMap and Multi-Temporal Satellite Data

S. Ghaffarian, N. Kerle, Edoardo Pasolli, Jamal Jokar Arsanjani

Research output: Contribution to journalArticleAcademicpeer-review

9 Downloads (Pure)

Abstract

First responders and recovery planners need accurate and quickly derived information about the status of buildings as well as newly built ones to both help victims and to make decisions for reconstruction processes after a disaster. Deep learning and, in particular, convolutional neural network (CNN)-based approaches have recently become state-of-the-art methods to extract information from remote sensing images, in particular for image-based structural damage assessment. However, they are predominantly based on manually extracted training samples. In the present study, we use pre-disaster OpenStreetMap building data to automatically generate training samples to train the proposed deep learning approach after the co-registration of the map and the satellite images. The proposed deep learning framework is based on the U-net design with residual connections, which has been shown to be an effective method to increase the efficiency of CNN-based models. The ResUnet is followed by a Conditional Random Field (CRF) implementation to further refine the results. Experimental analysis was carried out on selected very high resolution (VHR) satellite images representing various scenarios after the 2013 Super Typhoon Haiyan in both the damage and the recovery phases in Tacloban, the Philippines. The results show the robustness of the proposed ResUnet-CRF framework in updating the building map after a disaster for both damage and recovery situations by producing an overall F1-score of 84.2%.
Original languageEnglish
Article number2427
Pages (from-to)1-20
Number of pages20
JournalRemote sensing
Volume11
Issue number20
DOIs
Publication statusPublished - 19 Oct 2019

Fingerprint

satellite data
disaster
learning
damage
typhoon
train
remote sensing
method
satellite image
analysis
decision
registration
need
damage assessment
state of the art

Keywords

  • post-disaster
  • building database update
  • damage assessment
  • recovery asssessment
  • OpenStreetMap (OSM)
  • Deep learning
  • Convolutional Neural Network (CNN)
  • multi-temporal satellite imagery
  • U-Net
  • super typhoon haiyan
  • ITC-ISI-JOURNAL-ARTICLE
  • ITC-GOLD

Cite this

@article{aeb6a188d4b14a4a8758c0cce605fba8,
title = "Post-Disaster Building Database Updating Using Automated Deep Learning: An Integration of Pre-Disaster OpenStreetMap and Multi-Temporal Satellite Data",
abstract = "First responders and recovery planners need accurate and quickly derived information about the status of buildings as well as newly built ones to both help victims and to make decisions for reconstruction processes after a disaster. Deep learning and, in particular, convolutional neural network (CNN)-based approaches have recently become state-of-the-art methods to extract information from remote sensing images, in particular for image-based structural damage assessment. However, they are predominantly based on manually extracted training samples. In the present study, we use pre-disaster OpenStreetMap building data to automatically generate training samples to train the proposed deep learning approach after the co-registration of the map and the satellite images. The proposed deep learning framework is based on the U-net design with residual connections, which has been shown to be an effective method to increase the efficiency of CNN-based models. The ResUnet is followed by a Conditional Random Field (CRF) implementation to further refine the results. Experimental analysis was carried out on selected very high resolution (VHR) satellite images representing various scenarios after the 2013 Super Typhoon Haiyan in both the damage and the recovery phases in Tacloban, the Philippines. The results show the robustness of the proposed ResUnet-CRF framework in updating the building map after a disaster for both damage and recovery situations by producing an overall F1-score of 84.2{\%}.",
keywords = "post-disaster, building database update, damage assessment, recovery asssessment, OpenStreetMap (OSM), Deep learning, Convolutional Neural Network (CNN), multi-temporal satellite imagery, U-Net, super typhoon haiyan, ITC-ISI-JOURNAL-ARTICLE, ITC-GOLD",
author = "S. Ghaffarian and N. Kerle and Edoardo Pasolli and {Jokar Arsanjani}, Jamal",
year = "2019",
month = "10",
day = "19",
doi = "10.3390/rs11202427",
language = "English",
volume = "11",
pages = "1--20",
journal = "Remote sensing",
issn = "2072-4292",
publisher = "Multidisciplinary Digital Publishing Institute",
number = "20",

}

Post-Disaster Building Database Updating Using Automated Deep Learning : An Integration of Pre-Disaster OpenStreetMap and Multi-Temporal Satellite Data. / Ghaffarian, S.; Kerle, N.; Pasolli, Edoardo; Jokar Arsanjani, Jamal.

In: Remote sensing, Vol. 11, No. 20, 2427, 19.10.2019, p. 1-20.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Post-Disaster Building Database Updating Using Automated Deep Learning

T2 - An Integration of Pre-Disaster OpenStreetMap and Multi-Temporal Satellite Data

AU - Ghaffarian, S.

AU - Kerle, N.

AU - Pasolli, Edoardo

AU - Jokar Arsanjani, Jamal

PY - 2019/10/19

Y1 - 2019/10/19

N2 - First responders and recovery planners need accurate and quickly derived information about the status of buildings as well as newly built ones to both help victims and to make decisions for reconstruction processes after a disaster. Deep learning and, in particular, convolutional neural network (CNN)-based approaches have recently become state-of-the-art methods to extract information from remote sensing images, in particular for image-based structural damage assessment. However, they are predominantly based on manually extracted training samples. In the present study, we use pre-disaster OpenStreetMap building data to automatically generate training samples to train the proposed deep learning approach after the co-registration of the map and the satellite images. The proposed deep learning framework is based on the U-net design with residual connections, which has been shown to be an effective method to increase the efficiency of CNN-based models. The ResUnet is followed by a Conditional Random Field (CRF) implementation to further refine the results. Experimental analysis was carried out on selected very high resolution (VHR) satellite images representing various scenarios after the 2013 Super Typhoon Haiyan in both the damage and the recovery phases in Tacloban, the Philippines. The results show the robustness of the proposed ResUnet-CRF framework in updating the building map after a disaster for both damage and recovery situations by producing an overall F1-score of 84.2%.

AB - First responders and recovery planners need accurate and quickly derived information about the status of buildings as well as newly built ones to both help victims and to make decisions for reconstruction processes after a disaster. Deep learning and, in particular, convolutional neural network (CNN)-based approaches have recently become state-of-the-art methods to extract information from remote sensing images, in particular for image-based structural damage assessment. However, they are predominantly based on manually extracted training samples. In the present study, we use pre-disaster OpenStreetMap building data to automatically generate training samples to train the proposed deep learning approach after the co-registration of the map and the satellite images. The proposed deep learning framework is based on the U-net design with residual connections, which has been shown to be an effective method to increase the efficiency of CNN-based models. The ResUnet is followed by a Conditional Random Field (CRF) implementation to further refine the results. Experimental analysis was carried out on selected very high resolution (VHR) satellite images representing various scenarios after the 2013 Super Typhoon Haiyan in both the damage and the recovery phases in Tacloban, the Philippines. The results show the robustness of the proposed ResUnet-CRF framework in updating the building map after a disaster for both damage and recovery situations by producing an overall F1-score of 84.2%.

KW - post-disaster

KW - building database update

KW - damage assessment

KW - recovery asssessment

KW - OpenStreetMap (OSM)

KW - Deep learning

KW - Convolutional Neural Network (CNN)

KW - multi-temporal satellite imagery

KW - U-Net

KW - super typhoon haiyan

KW - ITC-ISI-JOURNAL-ARTICLE

KW - ITC-GOLD

UR - https://ezproxy2.utwente.nl/login?url=https://library.itc.utwente.nl/login/2019/isi/ghaffarian_pos.pdf

U2 - 10.3390/rs11202427

DO - 10.3390/rs11202427

M3 - Article

VL - 11

SP - 1

EP - 20

JO - Remote sensing

JF - Remote sensing

SN - 2072-4292

IS - 20

M1 - 2427

ER -