超越漂亮图片：Sentinel2图像的单图像和多图像超分辨率组合（英）

COMBINED SINGLE- AND MULTI-IMAGESUPER-RESOLUTION FOR SENTINEL-2 IMAGES Aditya Retnanto, Son H. Le, Sebastian Mueller, Armin Leitner, Michael Riffler,Konrad Schindler, and Yohan Iddawela ADB ECONOMICSWORKING PAPER SERIES Beyond Pretty Pictures: Combined Single- and Multi-ImageSuper-Resolution for Sentinel-2 Images Aditya Retnanto (aretnanto.consultant@adb.org),Son H. Le (shle.consultant@adb.org), and SebastianMueller (smueller.consultant@adb.org) are consultantsand Yohan Iddawela (yiddawela@adb.org) is aneconomist (data science) at the Economic Researchand Development Impact Department, AsianDevelopment Bank. Armin Leitner (leitner@geoville.com)is a data scientist and Michael Riffler (riffler@geoville.com)is head of research and innovation at GeoVilleInformation Systems and Data Processing GmbH.Konrad Schindler (schindler@ethz.ch) is a fullprofessor and head of the Institute of Geodesy andPhotogrammetry at ETH Zürich. Aditya Retnanto, Son H. Le, Sebastian Mueller,Armin Leitner, Michael Riffler, Konrad Schindler,and Yohan Iddawela No. 823 | November 2025 TheADB Economics Working Paper Seriespresents research in progress to elicit commentsand encourage debate on development issuesin Asia and the Pacific. The views expressedare those of the authors and do not necessarilyreflect the views and policies of ADB orits Board of Governors or the governmentsthey represent. Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) © 2025 Asian Development Bank6 ADB Avenue, Mandaluyong City, 1550 Metro Manila, PhilippinesTel +63 2 8632 4444; Fax +63 2 8636 2444www.adb.org Some rights reserved. Published in 2025. ISSN 2313-6537 (print), 2313-6545 (PDF)Publication Stock No. WPS250476-2DOI: http://dx.doi.org/10.22617/WPS250476-2 The views expressed in this publication are those of the authors and do not necessarily reflect the views and policiesof the Asian Development Bank (ADB) or its Board of Governors or the governments they represent. ADB does not guarantee the accuracy of the data included in this publication and accepts no responsibility for anyconsequence of their use. The mention of specific companies or products of manufacturers does not imply that theyare endorsed or recommended by ADB in preference to others of a similar nature that are not mentioned. By making any designation of or reference to a particular territory or geographic area in this document, ADB does notintend to make any judgments as to the legal or other status of any territory or area. This publication is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO)https://creativecommons.org/licenses/by/3.0/igo/. By using the content of this publication, you agree to be boundby the terms of this license. For attribution, translations, adaptations, and permissions, please read the provisionsand terms of use at https://www.adb.org/terms-use#openaccess. This CC license does not apply to non-ADB copyright materials in this publication. If the material is attributedto another source, please contact the copyright owner or publisher of that source for permission to reproduce it.ADB cannot be held liable for any claims that arise as a result of your use of the material. Please contact pubsmarketing@adb.org if you have questions or comments with respect to content, or if you wishto obtain copyright permission for your intended use that does not fall within these terms, or for permission to usethe ADB logo. ABSTRACT Super-resolution (SR) aims to increase the resolution of satellite images by reconstructinghigh-frequency details, which go beyond naive upsampling. This has particular relevancefor Earth observation missions like Sentinel-2, which offer frequent, regular coverage atno cost, but at coarse resolution. The pixel footprint of Sentinel-2 coverage is too large tocapture small features like houses, streets, or hedge rows. To address this, we presentSEN4X, a hybrid super-resolution architecture that combines the advantages of single-image and multi-image techniques. It combines temporal oversampling from repeatedSentinel-2 acquisitions with a learned prior from high-resolution Pléiades Neo data. Indoing so, SEN4X upgrades Sentinel-2 imagery to 2.5 m ground sampling distance. Wetest the super-resolved images on urban land-cover classification in Ha Noi, Viet Nam.We find that they lead to significant performance improvement over state-of-the-art super-resolution baselines. Keywords:super-resolution, remote sensing, Sentinel-2, land-cover classificationJEL code:Y8 1INTRODUCTION In many development contexts, timely and cost-effective access to high-resolutiongeospatialinformation is essential for planning,monitoring,and evaluation.Fromidentifying informal settlements and assessing disaster impacts to tracking deforestationor urban sprawl, detailed satellite imagery plays a critical role. However, truly highresolution (HR) satellite data is expensive and coverage is sparse, especially in the GlobalSouth. In cont