Satellite Image Classification Of Building Damages Using Airborne And Satellite Image Samples In A Deep Learning Approach

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

8 Citations (Scopus)
181 Downloads (Pure)

Abstract

The localization and detailed assessment of damaged buildings after a disastrous event is of utmost importance to guide response operations, recovery tasks or for insurance purposes. Several remote sensing platforms and sensors are currently used for the manual detection of building damages. However, there is an overall interest in the use of automated methods to perform this task, regardless of the used platform. Owing to its synoptic coverage and predictable availability, satellite imagery is currently used as input for the identification of building damages by the International Charter, as well as the Copernicus Emergency Management Service for the production of damage grading and reference maps. Recently proposed methods to perform image classification of building damages rely on convolutional neural networks (CNN). These are usually trained with only satellite image samples in a binary classification problem, however the number of samples derived from these images is often limited, affecting the quality of the classification results. The use of up/down-sampling image samples during the training of a CNN, has demonstrated to improve several image recognition tasks in remote sensing. However, it is currently unclear if this multi resolution information can also be captured from images with different spatial resolutions like satellite and airborne imagery (from both manned and unmanned platforms). In this paper, a CNN framework using residual connections and dilated convolutions is used considering both manned and unmanned aerial image samples to perform the satellite image classification of building damages. Three network configurations, trained with multi-resolution image samples are compared against two benchmark networks where only satellite image samples are used. Combining feature maps generated from airborne and satellite image samples, and refining these using only the satellite image samples, improved nearly 4% the overall satellite image classification of building damages.
Original languageEnglish
Title of host publicationISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Subtitle of host publicationISPRS TC II Mid-term Symposium “Towards Photogrammetry 2020”
PublisherInternational Society for Photogrammetry and Remote Sensing (ISPRS)
Pages89-96
Number of pages8
VolumeIV
Edition2
DOIs
Publication statusPublished - 6 Jun 2018

Fingerprint

image classification
learning
damage
satellite imagery
airborne sensing
remote sensing
image resolution
satellite image
spatial resolution
sensor
sampling

Keywords

  • ITC-GOLD

Cite this

Duarte, D., Nex, F. C., Kerle, N., & Vosselman, G. (2018). Satellite Image Classification Of Building Damages Using Airborne And Satellite Image Samples In A Deep Learning Approach. In ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences: ISPRS TC II Mid-term Symposium “Towards Photogrammetry 2020” (2 ed., Vol. IV, pp. 89-96). International Society for Photogrammetry and Remote Sensing (ISPRS). https://doi.org/10.5194/isprs-annals-IV-2-89-2018
Duarte, D. ; Nex, F.C. ; Kerle, N. ; Vosselman, G. / Satellite Image Classification Of Building Damages Using Airborne And Satellite Image Samples In A Deep Learning Approach. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences: ISPRS TC II Mid-term Symposium “Towards Photogrammetry 2020”. Vol. IV 2. ed. International Society for Photogrammetry and Remote Sensing (ISPRS), 2018. pp. 89-96
@inbook{f4da1bb8dd1346da945fc7edf13d8fdc,
title = "Satellite Image Classification Of Building Damages Using Airborne And Satellite Image Samples In A Deep Learning Approach",
abstract = "The localization and detailed assessment of damaged buildings after a disastrous event is of utmost importance to guide response operations, recovery tasks or for insurance purposes. Several remote sensing platforms and sensors are currently used for the manual detection of building damages. However, there is an overall interest in the use of automated methods to perform this task, regardless of the used platform. Owing to its synoptic coverage and predictable availability, satellite imagery is currently used as input for the identification of building damages by the International Charter, as well as the Copernicus Emergency Management Service for the production of damage grading and reference maps. Recently proposed methods to perform image classification of building damages rely on convolutional neural networks (CNN). These are usually trained with only satellite image samples in a binary classification problem, however the number of samples derived from these images is often limited, affecting the quality of the classification results. The use of up/down-sampling image samples during the training of a CNN, has demonstrated to improve several image recognition tasks in remote sensing. However, it is currently unclear if this multi resolution information can also be captured from images with different spatial resolutions like satellite and airborne imagery (from both manned and unmanned platforms). In this paper, a CNN framework using residual connections and dilated convolutions is used considering both manned and unmanned aerial image samples to perform the satellite image classification of building damages. Three network configurations, trained with multi-resolution image samples are compared against two benchmark networks where only satellite image samples are used. Combining feature maps generated from airborne and satellite image samples, and refining these using only the satellite image samples, improved nearly 4{\%} the overall satellite image classification of building damages.",
keywords = "ITC-GOLD",
author = "D. Duarte and F.C. Nex and N. Kerle and G. Vosselman",
year = "2018",
month = "6",
day = "6",
doi = "10.5194/isprs-annals-IV-2-89-2018",
language = "English",
volume = "IV",
pages = "89--96",
booktitle = "ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences",
publisher = "International Society for Photogrammetry and Remote Sensing (ISPRS)",
edition = "2",

}

Duarte, D, Nex, FC, Kerle, N & Vosselman, G 2018, Satellite Image Classification Of Building Damages Using Airborne And Satellite Image Samples In A Deep Learning Approach. in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences: ISPRS TC II Mid-term Symposium “Towards Photogrammetry 2020”. 2 edn, vol. IV, International Society for Photogrammetry and Remote Sensing (ISPRS), pp. 89-96. https://doi.org/10.5194/isprs-annals-IV-2-89-2018

Satellite Image Classification Of Building Damages Using Airborne And Satellite Image Samples In A Deep Learning Approach. / Duarte, D.; Nex, F.C.; Kerle, N.; Vosselman, G.

ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences: ISPRS TC II Mid-term Symposium “Towards Photogrammetry 2020”. Vol. IV 2. ed. International Society for Photogrammetry and Remote Sensing (ISPRS), 2018. p. 89-96.

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

TY - CHAP

T1 - Satellite Image Classification Of Building Damages Using Airborne And Satellite Image Samples In A Deep Learning Approach

AU - Duarte, D.

AU - Nex, F.C.

AU - Kerle, N.

AU - Vosselman, G.

PY - 2018/6/6

Y1 - 2018/6/6

N2 - The localization and detailed assessment of damaged buildings after a disastrous event is of utmost importance to guide response operations, recovery tasks or for insurance purposes. Several remote sensing platforms and sensors are currently used for the manual detection of building damages. However, there is an overall interest in the use of automated methods to perform this task, regardless of the used platform. Owing to its synoptic coverage and predictable availability, satellite imagery is currently used as input for the identification of building damages by the International Charter, as well as the Copernicus Emergency Management Service for the production of damage grading and reference maps. Recently proposed methods to perform image classification of building damages rely on convolutional neural networks (CNN). These are usually trained with only satellite image samples in a binary classification problem, however the number of samples derived from these images is often limited, affecting the quality of the classification results. The use of up/down-sampling image samples during the training of a CNN, has demonstrated to improve several image recognition tasks in remote sensing. However, it is currently unclear if this multi resolution information can also be captured from images with different spatial resolutions like satellite and airborne imagery (from both manned and unmanned platforms). In this paper, a CNN framework using residual connections and dilated convolutions is used considering both manned and unmanned aerial image samples to perform the satellite image classification of building damages. Three network configurations, trained with multi-resolution image samples are compared against two benchmark networks where only satellite image samples are used. Combining feature maps generated from airborne and satellite image samples, and refining these using only the satellite image samples, improved nearly 4% the overall satellite image classification of building damages.

AB - The localization and detailed assessment of damaged buildings after a disastrous event is of utmost importance to guide response operations, recovery tasks or for insurance purposes. Several remote sensing platforms and sensors are currently used for the manual detection of building damages. However, there is an overall interest in the use of automated methods to perform this task, regardless of the used platform. Owing to its synoptic coverage and predictable availability, satellite imagery is currently used as input for the identification of building damages by the International Charter, as well as the Copernicus Emergency Management Service for the production of damage grading and reference maps. Recently proposed methods to perform image classification of building damages rely on convolutional neural networks (CNN). These are usually trained with only satellite image samples in a binary classification problem, however the number of samples derived from these images is often limited, affecting the quality of the classification results. The use of up/down-sampling image samples during the training of a CNN, has demonstrated to improve several image recognition tasks in remote sensing. However, it is currently unclear if this multi resolution information can also be captured from images with different spatial resolutions like satellite and airborne imagery (from both manned and unmanned platforms). In this paper, a CNN framework using residual connections and dilated convolutions is used considering both manned and unmanned aerial image samples to perform the satellite image classification of building damages. Three network configurations, trained with multi-resolution image samples are compared against two benchmark networks where only satellite image samples are used. Combining feature maps generated from airborne and satellite image samples, and refining these using only the satellite image samples, improved nearly 4% the overall satellite image classification of building damages.

KW - ITC-GOLD

UR - https://ezproxy2.utwente.nl/login?url=https://webapps.itc.utwente.nl/library/2018/chap/kerle_sat.pdf

U2 - 10.5194/isprs-annals-IV-2-89-2018

DO - 10.5194/isprs-annals-IV-2-89-2018

M3 - Chapter

VL - IV

SP - 89

EP - 96

BT - ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences

PB - International Society for Photogrammetry and Remote Sensing (ISPRS)

ER -

Duarte D, Nex FC, Kerle N, Vosselman G. Satellite Image Classification Of Building Damages Using Airborne And Satellite Image Samples In A Deep Learning Approach. In ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences: ISPRS TC II Mid-term Symposium “Towards Photogrammetry 2020”. 2 ed. Vol. IV. International Society for Photogrammetry and Remote Sensing (ISPRS). 2018. p. 89-96 https://doi.org/10.5194/isprs-annals-IV-2-89-2018