Fusion of Aerial and Satellite Images for Automatic Extraction of Building Footprint Information Using Deep Neural Networks

The analysis of aerial and satellite images for building footprint detection is one of the major challenges in photogrammetry and remote sensing. This information is useful for various applications, such as urban planning, disaster monitoring, and 3D city modeling. However, it has become a significa...

Full description

Saved in:
Bibliographic Details
Main Authors: Ehsan Haghighi Gashti, Hanieh Bahiraei, Mohammad Javad Valadan Zoej, Ebrahim Ghaderpour
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Information
Subjects:
Online Access:https://www.mdpi.com/2078-2489/16/5/380
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850257214424481792
author Ehsan Haghighi Gashti
Hanieh Bahiraei
Mohammad Javad Valadan Zoej
Ebrahim Ghaderpour
author_facet Ehsan Haghighi Gashti
Hanieh Bahiraei
Mohammad Javad Valadan Zoej
Ebrahim Ghaderpour
author_sort Ehsan Haghighi Gashti
collection DOAJ
description The analysis of aerial and satellite images for building footprint detection is one of the major challenges in photogrammetry and remote sensing. This information is useful for various applications, such as urban planning, disaster monitoring, and 3D city modeling. However, it has become a significant challenge due to the diverse characteristics of buildings, such as shape, size, and shadow interference. This study investigated the simultaneous use of aerial and satellite images to improve the accuracy of deep learning models in building footprint detection. For this purpose, aerial images with a spatial resolution of 30 cm and Sentinel-2 satellite imagery were employed. Several satellite-derived spectral indices were extracted from the Sentinel-2 image. Then, U-Net models combined with ResNet-18 and ResNet-34 were trained on these data. The results showed that the combination of the U-Net model with ResNet-34, trained on a dataset obtained by integrating aerial images and satellite indices, referred to as RGB–Sentinel–ResNet34, achieved the best performance among the evaluated models. This model attained an accuracy of 96.99%, an F1-score of 90.57%, and an Intersection over Union of 73.86%. Compared to other models, RGB–Sentinel–ResNet34 showed a significant improvement in accuracy and generalization capability. The findings indicated that the simultaneous use of aerial and satellite data can substantially enhance the accuracy of building footprint detection.
format Article
id doaj-art-fb302e2331be4da3a7dbb42d4408c8e6
institution OA Journals
issn 2078-2489
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Information
spelling doaj-art-fb302e2331be4da3a7dbb42d4408c8e62025-08-20T01:56:28ZengMDPI AGInformation2078-24892025-05-0116538010.3390/info16050380Fusion of Aerial and Satellite Images for Automatic Extraction of Building Footprint Information Using Deep Neural NetworksEhsan Haghighi Gashti0Hanieh Bahiraei1Mohammad Javad Valadan Zoej2Ebrahim Ghaderpour3School of Surveying and Geospatial Eng, College of Engineering, University of Tehran, 1417935840 Tehran, IranFaculty of Geomatics Engineering, K. N. Toosi University of Technology, 1969764499 Tehran, IranFaculty of Geomatics Engineering, K. N. Toosi University of Technology, 1969764499 Tehran, IranDepartment of Earth Sciences, Sapienza University of Rome, P. le Aldo Moro, 5, 00185 Rome, ItalyThe analysis of aerial and satellite images for building footprint detection is one of the major challenges in photogrammetry and remote sensing. This information is useful for various applications, such as urban planning, disaster monitoring, and 3D city modeling. However, it has become a significant challenge due to the diverse characteristics of buildings, such as shape, size, and shadow interference. This study investigated the simultaneous use of aerial and satellite images to improve the accuracy of deep learning models in building footprint detection. For this purpose, aerial images with a spatial resolution of 30 cm and Sentinel-2 satellite imagery were employed. Several satellite-derived spectral indices were extracted from the Sentinel-2 image. Then, U-Net models combined with ResNet-18 and ResNet-34 were trained on these data. The results showed that the combination of the U-Net model with ResNet-34, trained on a dataset obtained by integrating aerial images and satellite indices, referred to as RGB–Sentinel–ResNet34, achieved the best performance among the evaluated models. This model attained an accuracy of 96.99%, an F1-score of 90.57%, and an Intersection over Union of 73.86%. Compared to other models, RGB–Sentinel–ResNet34 showed a significant improvement in accuracy and generalization capability. The findings indicated that the simultaneous use of aerial and satellite data can substantially enhance the accuracy of building footprint detection.https://www.mdpi.com/2078-2489/16/5/380deep learningsemantic segmentationU-NetResNetSentinel-2
spellingShingle Ehsan Haghighi Gashti
Hanieh Bahiraei
Mohammad Javad Valadan Zoej
Ebrahim Ghaderpour
Fusion of Aerial and Satellite Images for Automatic Extraction of Building Footprint Information Using Deep Neural Networks
Information
deep learning
semantic segmentation
U-Net
ResNet
Sentinel-2
title Fusion of Aerial and Satellite Images for Automatic Extraction of Building Footprint Information Using Deep Neural Networks
title_full Fusion of Aerial and Satellite Images for Automatic Extraction of Building Footprint Information Using Deep Neural Networks
title_fullStr Fusion of Aerial and Satellite Images for Automatic Extraction of Building Footprint Information Using Deep Neural Networks
title_full_unstemmed Fusion of Aerial and Satellite Images for Automatic Extraction of Building Footprint Information Using Deep Neural Networks
title_short Fusion of Aerial and Satellite Images for Automatic Extraction of Building Footprint Information Using Deep Neural Networks
title_sort fusion of aerial and satellite images for automatic extraction of building footprint information using deep neural networks
topic deep learning
semantic segmentation
U-Net
ResNet
Sentinel-2
url https://www.mdpi.com/2078-2489/16/5/380
work_keys_str_mv AT ehsanhaghighigashti fusionofaerialandsatelliteimagesforautomaticextractionofbuildingfootprintinformationusingdeepneuralnetworks
AT haniehbahiraei fusionofaerialandsatelliteimagesforautomaticextractionofbuildingfootprintinformationusingdeepneuralnetworks
AT mohammadjavadvaladanzoej fusionofaerialandsatelliteimagesforautomaticextractionofbuildingfootprintinformationusingdeepneuralnetworks
AT ebrahimghaderpour fusionofaerialandsatelliteimagesforautomaticextractionofbuildingfootprintinformationusingdeepneuralnetworks