How high are we? Large-scale building height estimation at 10 m using Sentinel-1 SAR and Sentinel-2 MSI time series

IF 11.1 1区 地球科学 Q1 ENVIRONMENTAL SCIENCES Remote Sensing of Environment Pub Date : 2024-12-16 DOI:10.1016/j.rse.2024.114556
Ritu Yadav, Andrea Nascetti, Yifang Ban
{"title":"How high are we? Large-scale building height estimation at 10 m using Sentinel-1 SAR and Sentinel-2 MSI time series","authors":"Ritu Yadav, Andrea Nascetti, Yifang Ban","doi":"10.1016/j.rse.2024.114556","DOIUrl":null,"url":null,"abstract":"Accurate building height estimation is essential to support urbanization monitoring, environmental impact analysis and sustainable urban planning. However, conducting large-scale building height estimation remains a significant challenge. While deep learning (DL) has proven effective for large-scale mapping tasks, there is a lack of advanced DL models specifically tailored for height estimation, particularly when using open-source Earth observation data. In this study, we propose T-SwinUNet, an advanced DL model for large-scale building height estimation leveraging Sentinel-1 SAR and Sentinel-2 multispectral time series. T-SwinUNet model contains a feature extractor with local/global feature comprehension capabilities, a temporal attention module to learn the correlation between constant and variable features of building objects over time and an efficient multitask decoder to predict building height at 10 m spatial resolution. The model is trained and evaluated on data from the Netherlands, Switzerland, Estonia, and Germany, and its generalizability is evaluated on an out-of-distribution (OOD) test set from ten additional cities from other European countries. Our study incorporates extensive model evaluations, ablation experiments, and comparisons with established models. T-SwinUNet predicts building height with a Root Mean Square Error (RMSE) of 1.89 m, outperforming state-of-the-art models at 10 m spatial resolution. Its strong generalization to the OOD test set (RMSE of 3.2 m) underscores its potential for low-cost building height estimation across Europe, with future scalability to other regions. Furthermore, the assessment at 100 m resolution reveals that T-SwinUNet (0.29 m RMSE, 0.75 <span><span style=\"\"></span><span data-mathml='&lt;math xmlns=\"http://www.w3.org/1998/Math/MathML\"&gt;&lt;msup is=\"true\"&gt;&lt;mrow is=\"true\"&gt;&lt;mi is=\"true\"&gt;R&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow is=\"true\"&gt;&lt;mn is=\"true\"&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/math&gt;' role=\"presentation\" style=\"font-size: 90%; display: inline-block; position: relative;\" tabindex=\"0\"><svg aria-hidden=\"true\" focusable=\"false\" height=\"2.432ex\" role=\"img\" style=\"vertical-align: -0.235ex;\" viewbox=\"0 -945.9 1213.4 1047.3\" width=\"2.818ex\" xmlns:xlink=\"http://www.w3.org/1999/xlink\"><g fill=\"currentColor\" stroke=\"currentColor\" stroke-width=\"0\" transform=\"matrix(1 0 0 -1 0 0)\"><g is=\"true\"><g is=\"true\"><g is=\"true\"><use xlink:href=\"#MJMATHI-52\"></use></g></g><g is=\"true\" transform=\"translate(759,410)\"><g is=\"true\"><use transform=\"scale(0.707)\" xlink:href=\"#MJMAIN-32\"></use></g></g></g></g></svg><span role=\"presentation\"><math xmlns=\"http://www.w3.org/1998/Math/MathML\"><msup is=\"true\"><mrow is=\"true\"><mi is=\"true\">R</mi></mrow><mrow is=\"true\"><mn is=\"true\">2</mn></mrow></msup></math></span></span><script type=\"math/mml\"><math><msup is=\"true\"><mrow is=\"true\"><mi is=\"true\">R</mi></mrow><mrow is=\"true\"><mn is=\"true\">2</mn></mrow></msup></math></script></span>) also outperformed the global building height product GHSL-Built-H R2023A product(0.56 m RMSE and 0.37 <span><span style=\"\"></span><span data-mathml='&lt;math xmlns=\"http://www.w3.org/1998/Math/MathML\"&gt;&lt;msup is=\"true\"&gt;&lt;mrow is=\"true\"&gt;&lt;mi is=\"true\"&gt;R&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow is=\"true\"&gt;&lt;mn is=\"true\"&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/math&gt;' role=\"presentation\" style=\"font-size: 90%; display: inline-block; position: relative;\" tabindex=\"0\"><svg aria-hidden=\"true\" focusable=\"false\" height=\"2.432ex\" role=\"img\" style=\"vertical-align: -0.235ex;\" viewbox=\"0 -945.9 1213.4 1047.3\" width=\"2.818ex\" xmlns:xlink=\"http://www.w3.org/1999/xlink\"><g fill=\"currentColor\" stroke=\"currentColor\" stroke-width=\"0\" transform=\"matrix(1 0 0 -1 0 0)\"><g is=\"true\"><g is=\"true\"><g is=\"true\"><use xlink:href=\"#MJMATHI-52\"></use></g></g><g is=\"true\" transform=\"translate(759,410)\"><g is=\"true\"><use transform=\"scale(0.707)\" xlink:href=\"#MJMAIN-32\"></use></g></g></g></g></svg><span role=\"presentation\"><math xmlns=\"http://www.w3.org/1998/Math/MathML\"><msup is=\"true\"><mrow is=\"true\"><mi is=\"true\">R</mi></mrow><mrow is=\"true\"><mn is=\"true\">2</mn></mrow></msup></math></span></span><script type=\"math/mml\"><math><msup is=\"true\"><mrow is=\"true\"><mi is=\"true\">R</mi></mrow><mrow is=\"true\"><mn is=\"true\">2</mn></mrow></msup></math></script></span>). Our implementation is available at: <span><span>https://github.com/RituYadav92/Building-Height-Estimation</span><svg aria-label=\"Opens in new window\" focusable=\"false\" height=\"20\" viewbox=\"0 0 8 8\"><path d=\"M1.12949 2.1072V1H7V6.85795H5.89111V2.90281L0.784057 8L0 7.21635L5.11902 2.1072H1.12949Z\"></path></svg></span>.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"10 1","pages":""},"PeriodicalIF":11.1000,"publicationDate":"2024-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Remote Sensing of Environment","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1016/j.rse.2024.114556","RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0

Abstract

Accurate building height estimation is essential to support urbanization monitoring, environmental impact analysis and sustainable urban planning. However, conducting large-scale building height estimation remains a significant challenge. While deep learning (DL) has proven effective for large-scale mapping tasks, there is a lack of advanced DL models specifically tailored for height estimation, particularly when using open-source Earth observation data. In this study, we propose T-SwinUNet, an advanced DL model for large-scale building height estimation leveraging Sentinel-1 SAR and Sentinel-2 multispectral time series. T-SwinUNet model contains a feature extractor with local/global feature comprehension capabilities, a temporal attention module to learn the correlation between constant and variable features of building objects over time and an efficient multitask decoder to predict building height at 10 m spatial resolution. The model is trained and evaluated on data from the Netherlands, Switzerland, Estonia, and Germany, and its generalizability is evaluated on an out-of-distribution (OOD) test set from ten additional cities from other European countries. Our study incorporates extensive model evaluations, ablation experiments, and comparisons with established models. T-SwinUNet predicts building height with a Root Mean Square Error (RMSE) of 1.89 m, outperforming state-of-the-art models at 10 m spatial resolution. Its strong generalization to the OOD test set (RMSE of 3.2 m) underscores its potential for low-cost building height estimation across Europe, with future scalability to other regions. Furthermore, the assessment at 100 m resolution reveals that T-SwinUNet (0.29 m RMSE, 0.75 R2) also outperformed the global building height product GHSL-Built-H R2023A product(0.56 m RMSE and 0.37 R2). Our implementation is available at: https://github.com/RituYadav92/Building-Height-Estimation.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
准确的建筑高度估算对于支持城市化监测、环境影响分析和可持续城市规划至关重要。然而,进行大规模建筑高度估算仍然是一项重大挑战。虽然深度学习(DL)已被证明对大规模测绘任务非常有效,但目前还缺乏专门用于高度估算的高级 DL 模型,尤其是在使用开源地球观测数据时。在本研究中,我们提出了 T-SwinUNet,一种利用 Sentinel-1 SAR 和 Sentinel-2 多光谱时间序列进行大规模建筑物高度估算的高级 DL 模型。T-SwinUNet 模型包含一个具有局部/全局特征理解能力的特征提取器、一个用于学习建筑物体随时间变化的恒定和可变特征之间的相关性的时间关注模块,以及一个用于预测 10 米空间分辨率下建筑高度的高效多任务解码器。该模型在荷兰、瑞士、爱沙尼亚和德国的数据上进行了训练和评估,并在来自其他欧洲国家另外十个城市的分布外(OOD)测试集上对其通用性进行了评估。我们的研究包括广泛的模型评估、消融实验以及与已有模型的比较。T-SwinUNet 预测建筑高度的均方根误差 (RMSE) 为 1.89 米,在 10 米空间分辨率下优于最先进的模型。它对 OOD 测试集(均方根误差为 3.2 米)的强大普适性强调了其在欧洲低成本建筑高度估算方面的潜力,未来还可扩展到其他地区。此外,在 100 米分辨率下进行的评估显示,T-SwinUNet(0.29 米均方根误差,0.75 R2R2)也优于全球建筑高度产品 GHSL-Built-H R2023A 产品(0.56 米均方根误差,0.37 R2R2)。我们的实施方案可在以下网址获取:https://github.com/RituYadav92/Building-Height-Estimation.
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Remote Sensing of Environment
Remote Sensing of Environment 环境科学-成像科学与照相技术
CiteScore
25.10
自引率
8.90%
发文量
455
审稿时长
53 days
期刊介绍: Remote Sensing of Environment (RSE) serves the Earth observation community by disseminating results on the theory, science, applications, and technology that contribute to advancing the field of remote sensing. With a thoroughly interdisciplinary approach, RSE encompasses terrestrial, oceanic, and atmospheric sensing. The journal emphasizes biophysical and quantitative approaches to remote sensing at local to global scales, covering a diverse range of applications and techniques. RSE serves as a vital platform for the exchange of knowledge and advancements in the dynamic field of remote sensing.
期刊最新文献
Ground surface displacement measurement from SAR imagery using deep learning Coupled hydrologic-electromagnetic framework to model permafrost active layer organic soil dielectric properties Joint mapping of melt pond bathymetry and water volume on sea ice using optical remote sensing images and physical reflectance models Quantitative characterization of global nighttime light: A method for measuring energy intensity based on radiant flux and SNPP-VIIRS data A flexible framework for built-up height mapping using ICESat-2 photons and multisource satellite observations
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1