Building change detection via semantic segmentation and difference extraction method

Siti Nor Khuzaimah Binti Amit, Shunta Saito, Yoshimitsu Aoki, Yasushi Kiyoki

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

Abstract

Google Earth with high-resolution imagery basically takes months to process new images before online updates. It is considered as a time consuming and slow process especially for post-disaster application. In this study, we aim to develop a fast and accurate method of updating maps by detecting local differences occurred over different time series; where only region with differences will be updated. In our system, aerial imageries from Massachusetts's building open datasets are used as training datasets; meanwhile Saitama district datasets are used as input images. Semantic segmentation is then applied to input images to get predicted map patches of building. Semantic segmentation is a pixel-wise classification of images by implementing convolutional neural network technique. Convolutional neural network technique is implemented due to being not only efficient in learning highly discriminative image features such as buildings, but also partially robust to incomplete and poorly registered target maps. Next, in order to understand overall changes occurred in an area, both semantic segmented images from the same scene are undergone change detection method. Lastly, difference extraction method is implemented to specify the category of building changes. The results reveal that our proposed method is able to overcome current time-consuming map updating problem. Hence map updating will be cheaper, faster and more effective especially post-disaster application, by leaving unchanged region and only updating changed region.

Original languageEnglish
Title of host publicationInformation Modelling and Knowledge Bases XXVIII
PublisherIOS Press
Pages249-257
Number of pages9
Volume292
ISBN (Electronic)9781614997191
DOIs
Publication statusPublished - 2017

Publication series

NameFrontiers in Artificial Intelligence and Applications
Volume292
ISSN (Print)09226389

Fingerprint

Semantics
Disasters
Neural networks
Time series
Pixels
Earth (planet)
Antennas

Keywords

  • aerial imagery
  • building change detection
  • convolutional neural network
  • difference extraction
  • semantic segmentation

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Amit, S. N. K. B., Saito, S., Aoki, Y., & Kiyoki, Y. (2017). Building change detection via semantic segmentation and difference extraction method. In Information Modelling and Knowledge Bases XXVIII (Vol. 292, pp. 249-257). (Frontiers in Artificial Intelligence and Applications; Vol. 292). IOS Press. https://doi.org/10.3233/978-1-61499-720-7-249

Building change detection via semantic segmentation and difference extraction method. / Amit, Siti Nor Khuzaimah Binti; Saito, Shunta; Aoki, Yoshimitsu; Kiyoki, Yasushi.

Information Modelling and Knowledge Bases XXVIII. Vol. 292 IOS Press, 2017. p. 249-257 (Frontiers in Artificial Intelligence and Applications; Vol. 292).

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

Amit, SNKB, Saito, S, Aoki, Y & Kiyoki, Y 2017, Building change detection via semantic segmentation and difference extraction method. in Information Modelling and Knowledge Bases XXVIII. vol. 292, Frontiers in Artificial Intelligence and Applications, vol. 292, IOS Press, pp. 249-257. https://doi.org/10.3233/978-1-61499-720-7-249
Amit SNKB, Saito S, Aoki Y, Kiyoki Y. Building change detection via semantic segmentation and difference extraction method. In Information Modelling and Knowledge Bases XXVIII. Vol. 292. IOS Press. 2017. p. 249-257. (Frontiers in Artificial Intelligence and Applications). https://doi.org/10.3233/978-1-61499-720-7-249
Amit, Siti Nor Khuzaimah Binti ; Saito, Shunta ; Aoki, Yoshimitsu ; Kiyoki, Yasushi. / Building change detection via semantic segmentation and difference extraction method. Information Modelling and Knowledge Bases XXVIII. Vol. 292 IOS Press, 2017. pp. 249-257 (Frontiers in Artificial Intelligence and Applications).
@inproceedings{f40f70588900437c8961607539d4b76b,
title = "Building change detection via semantic segmentation and difference extraction method",
abstract = "Google Earth with high-resolution imagery basically takes months to process new images before online updates. It is considered as a time consuming and slow process especially for post-disaster application. In this study, we aim to develop a fast and accurate method of updating maps by detecting local differences occurred over different time series; where only region with differences will be updated. In our system, aerial imageries from Massachusetts's building open datasets are used as training datasets; meanwhile Saitama district datasets are used as input images. Semantic segmentation is then applied to input images to get predicted map patches of building. Semantic segmentation is a pixel-wise classification of images by implementing convolutional neural network technique. Convolutional neural network technique is implemented due to being not only efficient in learning highly discriminative image features such as buildings, but also partially robust to incomplete and poorly registered target maps. Next, in order to understand overall changes occurred in an area, both semantic segmented images from the same scene are undergone change detection method. Lastly, difference extraction method is implemented to specify the category of building changes. The results reveal that our proposed method is able to overcome current time-consuming map updating problem. Hence map updating will be cheaper, faster and more effective especially post-disaster application, by leaving unchanged region and only updating changed region.",
keywords = "aerial imagery, building change detection, convolutional neural network, difference extraction, semantic segmentation",
author = "Amit, {Siti Nor Khuzaimah Binti} and Shunta Saito and Yoshimitsu Aoki and Yasushi Kiyoki",
year = "2017",
doi = "10.3233/978-1-61499-720-7-249",
language = "English",
volume = "292",
series = "Frontiers in Artificial Intelligence and Applications",
publisher = "IOS Press",
pages = "249--257",
booktitle = "Information Modelling and Knowledge Bases XXVIII",

}

TY - GEN

T1 - Building change detection via semantic segmentation and difference extraction method

AU - Amit, Siti Nor Khuzaimah Binti

AU - Saito, Shunta

AU - Aoki, Yoshimitsu

AU - Kiyoki, Yasushi

PY - 2017

Y1 - 2017

N2 - Google Earth with high-resolution imagery basically takes months to process new images before online updates. It is considered as a time consuming and slow process especially for post-disaster application. In this study, we aim to develop a fast and accurate method of updating maps by detecting local differences occurred over different time series; where only region with differences will be updated. In our system, aerial imageries from Massachusetts's building open datasets are used as training datasets; meanwhile Saitama district datasets are used as input images. Semantic segmentation is then applied to input images to get predicted map patches of building. Semantic segmentation is a pixel-wise classification of images by implementing convolutional neural network technique. Convolutional neural network technique is implemented due to being not only efficient in learning highly discriminative image features such as buildings, but also partially robust to incomplete and poorly registered target maps. Next, in order to understand overall changes occurred in an area, both semantic segmented images from the same scene are undergone change detection method. Lastly, difference extraction method is implemented to specify the category of building changes. The results reveal that our proposed method is able to overcome current time-consuming map updating problem. Hence map updating will be cheaper, faster and more effective especially post-disaster application, by leaving unchanged region and only updating changed region.

AB - Google Earth with high-resolution imagery basically takes months to process new images before online updates. It is considered as a time consuming and slow process especially for post-disaster application. In this study, we aim to develop a fast and accurate method of updating maps by detecting local differences occurred over different time series; where only region with differences will be updated. In our system, aerial imageries from Massachusetts's building open datasets are used as training datasets; meanwhile Saitama district datasets are used as input images. Semantic segmentation is then applied to input images to get predicted map patches of building. Semantic segmentation is a pixel-wise classification of images by implementing convolutional neural network technique. Convolutional neural network technique is implemented due to being not only efficient in learning highly discriminative image features such as buildings, but also partially robust to incomplete and poorly registered target maps. Next, in order to understand overall changes occurred in an area, both semantic segmented images from the same scene are undergone change detection method. Lastly, difference extraction method is implemented to specify the category of building changes. The results reveal that our proposed method is able to overcome current time-consuming map updating problem. Hence map updating will be cheaper, faster and more effective especially post-disaster application, by leaving unchanged region and only updating changed region.

KW - aerial imagery

KW - building change detection

KW - convolutional neural network

KW - difference extraction

KW - semantic segmentation

UR - http://www.scopus.com/inward/record.url?scp=85003025814&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85003025814&partnerID=8YFLogxK

U2 - 10.3233/978-1-61499-720-7-249

DO - 10.3233/978-1-61499-720-7-249

M3 - Conference contribution

AN - SCOPUS:85003025814

VL - 292

T3 - Frontiers in Artificial Intelligence and Applications

SP - 249

EP - 257

BT - Information Modelling and Knowledge Bases XXVIII

PB - IOS Press

ER -