Building Detection from Aerial Imagery using Inception Resnet UNET and UNET Models

Research output: Contribution to journalConference articleResearchpeer review

Authors

  • S. Aghayari
  • A. Hadavand
  • S. Mohamadnezhad Niazi
  • Mohammad Omidalizarandi

Research Organisations

External Research Organisations

  • Ideh Pardazan Tosseah Consulting Engineering Company
View graph of relations

Details

Original languageEnglish
Pages (from-to)9-17
Number of pages9
JournalISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Volume10
Issue number4/W1-2022
Publication statusPublished - 13 Jan 2023
Event6th SMPR and 4th GIResearch, ISPRS Geospatial Conference -
Duration: 19 Feb 202322 Feb 2023

Abstract

Buildings are one of the key components in change detection, urban planning, and monitoring. The automatic extraction of the building from high-resolution aerial imagery is still challenging due to the variations in their shapes, structures, textures, and colours. Recently, the convolutional neural networks (CNN) show a significant improvement in object detection and extraction that surpasses other methods. To extract building, in this paper two segmentation architectures, the UNet and the Inception ResNet UNet are implemented and then tested on the Inria aerial image datasets. The Inception ResNet UNet utilizes the Inception architecture and residual blocks. This makes the model wide and deep, though there are a few differences between numbers of UNet and Inception ResNet UNet parameters. The analyses show that UNet has a high rate of metrics in the training progress. However, on the unseen dataset, Inception ResNet UNet extracts buildings more accurately (97.95% accuracy and 0.96 in the dice metric) in comparison with UNet (94.30% accuracy and 0.55 in the dice metric).

Keywords

    Large-scale monitoring, Building detection, Image segmentation, Residual blocks, Skip connection

ASJC Scopus subject areas

Cite this

Building Detection from Aerial Imagery using Inception Resnet UNET and UNET Models. / Aghayari, S.; Hadavand, A.; Mohamadnezhad Niazi, S. et al.
In: ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. 10, No. 4/W1-2022, 13.01.2023, p. 9-17.

Research output: Contribution to journalConference articleResearchpeer review

Aghayari, S, Hadavand, A, Mohamadnezhad Niazi, S & Omidalizarandi, M 2023, 'Building Detection from Aerial Imagery using Inception Resnet UNET and UNET Models', ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. 10, no. 4/W1-2022, pp. 9-17. https://doi.org/10.5194/isprs-annals-X-4-W1-2022-9-2023
Aghayari, S., Hadavand, A., Mohamadnezhad Niazi, S., & Omidalizarandi, M. (2023). Building Detection from Aerial Imagery using Inception Resnet UNET and UNET Models. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 10(4/W1-2022), 9-17. https://doi.org/10.5194/isprs-annals-X-4-W1-2022-9-2023
Aghayari S, Hadavand A, Mohamadnezhad Niazi S, Omidalizarandi M. Building Detection from Aerial Imagery using Inception Resnet UNET and UNET Models. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences. 2023 Jan 13;10(4/W1-2022):9-17. doi: 10.5194/isprs-annals-X-4-W1-2022-9-2023
Aghayari, S. ; Hadavand, A. ; Mohamadnezhad Niazi, S. et al. / Building Detection from Aerial Imagery using Inception Resnet UNET and UNET Models. In: ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences. 2023 ; Vol. 10, No. 4/W1-2022. pp. 9-17.
Download
@article{7ca93f3965aa40748189d658ed0c9376,
title = "Building Detection from Aerial Imagery using Inception Resnet UNET and UNET Models",
abstract = "Buildings are one of the key components in change detection, urban planning, and monitoring. The automatic extraction of the building from high-resolution aerial imagery is still challenging due to the variations in their shapes, structures, textures, and colours. Recently, the convolutional neural networks (CNN) show a significant improvement in object detection and extraction that surpasses other methods. To extract building, in this paper two segmentation architectures, the UNet and the Inception ResNet UNet are implemented and then tested on the Inria aerial image datasets. The Inception ResNet UNet utilizes the Inception architecture and residual blocks. This makes the model wide and deep, though there are a few differences between numbers of UNet and Inception ResNet UNet parameters. The analyses show that UNet has a high rate of metrics in the training progress. However, on the unseen dataset, Inception ResNet UNet extracts buildings more accurately (97.95% accuracy and 0.96 in the dice metric) in comparison with UNet (94.30% accuracy and 0.55 in the dice metric).",
keywords = "Large-scale monitoring, Building detection, Image segmentation, Residual blocks, Skip connection",
author = "S. Aghayari and A. Hadavand and {Mohamadnezhad Niazi}, S. and Mohammad Omidalizarandi",
note = "Funding Information: This research was supported by Ideh Pardazan Tosseah Consulting Engineering Company. We are grateful to all of those with whom we have had the pleasure to work during this and other related projects.; 6th SMPR and 4th GIResearch, ISPRS Geospatial Conference ; Conference date: 19-02-2023 Through 22-02-2023",
year = "2023",
month = jan,
day = "13",
doi = "10.5194/isprs-annals-X-4-W1-2022-9-2023",
language = "English",
volume = "10",
pages = "9--17",
number = "4/W1-2022",

}

Download

TY - JOUR

T1 - Building Detection from Aerial Imagery using Inception Resnet UNET and UNET Models

AU - Aghayari, S.

AU - Hadavand, A.

AU - Mohamadnezhad Niazi, S.

AU - Omidalizarandi, Mohammad

N1 - Funding Information: This research was supported by Ideh Pardazan Tosseah Consulting Engineering Company. We are grateful to all of those with whom we have had the pleasure to work during this and other related projects.

PY - 2023/1/13

Y1 - 2023/1/13

N2 - Buildings are one of the key components in change detection, urban planning, and monitoring. The automatic extraction of the building from high-resolution aerial imagery is still challenging due to the variations in their shapes, structures, textures, and colours. Recently, the convolutional neural networks (CNN) show a significant improvement in object detection and extraction that surpasses other methods. To extract building, in this paper two segmentation architectures, the UNet and the Inception ResNet UNet are implemented and then tested on the Inria aerial image datasets. The Inception ResNet UNet utilizes the Inception architecture and residual blocks. This makes the model wide and deep, though there are a few differences between numbers of UNet and Inception ResNet UNet parameters. The analyses show that UNet has a high rate of metrics in the training progress. However, on the unseen dataset, Inception ResNet UNet extracts buildings more accurately (97.95% accuracy and 0.96 in the dice metric) in comparison with UNet (94.30% accuracy and 0.55 in the dice metric).

AB - Buildings are one of the key components in change detection, urban planning, and monitoring. The automatic extraction of the building from high-resolution aerial imagery is still challenging due to the variations in their shapes, structures, textures, and colours. Recently, the convolutional neural networks (CNN) show a significant improvement in object detection and extraction that surpasses other methods. To extract building, in this paper two segmentation architectures, the UNet and the Inception ResNet UNet are implemented and then tested on the Inria aerial image datasets. The Inception ResNet UNet utilizes the Inception architecture and residual blocks. This makes the model wide and deep, though there are a few differences between numbers of UNet and Inception ResNet UNet parameters. The analyses show that UNet has a high rate of metrics in the training progress. However, on the unseen dataset, Inception ResNet UNet extracts buildings more accurately (97.95% accuracy and 0.96 in the dice metric) in comparison with UNet (94.30% accuracy and 0.55 in the dice metric).

KW - Large-scale monitoring

KW - Building detection

KW - Image segmentation

KW - Residual blocks

KW - Skip connection

UR - http://www.scopus.com/inward/record.url?scp=85146937067&partnerID=8YFLogxK

U2 - 10.5194/isprs-annals-X-4-W1-2022-9-2023

DO - 10.5194/isprs-annals-X-4-W1-2022-9-2023

M3 - Conference article

VL - 10

SP - 9

EP - 17

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

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

SN - 2194-9042

IS - 4/W1-2022

T2 - 6th SMPR and 4th GIResearch, ISPRS Geospatial Conference

Y2 - 19 February 2023 through 22 February 2023

ER -