MADA-SSA: Multi-Axis Directional Attention and Spatial-Semantic Aggregation for Robust Drone-to-Satellite Geo-Localization

MADA-SSA: Multi-Axis Directional Attention and Spatial-Semantic Aggregation for Robust Drone-to-Satellite Geo-Localization | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article MADA-SSA: Multi-Axis Directional Attention and Spatial-Semantic Aggregation for Robust Drone-to-Satellite Geo-Localization He Xiao, Guang Yang, Jiaxing Liu, Pin Luo, Qiuming Liu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9512906/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Cross-view geo-localization serves as a key technique for drone autonomous nav- igation and disaster response. Existing methods sufer from limited receptive fields, weak directional geometric modeling, and unstable metric learning under extreme perspective changes. To address these challenges, we propose a unified framework that efectively integrates three core components: a Spatial-Semantic Multi-scale Aggregation (SSMA) module, which enhances global feature rep- resentation beyond local receptive fields; a Directional Strip Attention Block (DSAB) that refines geometric structure through multi-axis gating and diago- nal remapping to bridge directional perception gaps; and a Smooth Hard-Mining Loss (SHMLoss) that stabilizes metric learning with a density-aware weighting strategy. Experiments demonstrate that our method achieves 96.01% recall@1 and 92.84% AP on the University-1652 dataset, along with consistent perfor- mance across varying altitudes on SUES-200, establishing a reliable and robust paradigm for drone-to-satellite visual localization. The code and model files are publicly available at https://doi.org/10.5281/zenodo.19702776 . Cross-View geo-localization Drone to satellite Matching Strip Attention mechanism Metric Learning Multi-Scale Feature Fusion Full Text Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9512906","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":628753759,"identity":"8a02d35b-473e-468c-9d89-67d426c3e660","order_by":0,"name":"He Xiao","email":"","orcid":"","institution":"Jiangxi University of Science and Technology,Nanchang","correspondingAuthor":false,"prefix":"","firstName":"He","middleName":"","lastName":"Xiao","suffix":""},{"id":628753760,"identity":"7d05f099-9eec-4d6e-a9d1-d4c92b071a29","order_by":1,"name":"Guang Yang","email":"","orcid":"","institution":"Jiangxi University of Science and Technology,Nanchang","correspondingAuthor":false,"prefix":"","firstName":"Guang","middleName":"","lastName":"Yang","suffix":""},{"id":628753761,"identity":"830c18b9-0846-4c20-aa22-ccd7f07d381c","order_by":2,"name":"Jiaxing Liu","email":"","orcid":"","institution":"Jiangxi University of Science and Technology,Nanchang","correspondingAuthor":false,"prefix":"","firstName":"Jiaxing","middleName":"","lastName":"Liu","suffix":""},{"id":628753762,"identity":"103d3a2b-f019-4f89-9fac-07387bab1d42","order_by":3,"name":"Pin Luo","email":"","orcid":"","institution":"Jiangxi University of Science and Technology,Nanchang","correspondingAuthor":false,"prefix":"","firstName":"Pin","middleName":"","lastName":"Luo","suffix":""},{"id":628753763,"identity":"b3443600-a504-4815-8638-32a96d01abd8","order_by":4,"name":"Qiuming Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8ElEQVRIiWNgGAWjYBACPmYgkcDAwMPAwHzAACpogEcDAwMbQgtbAkipBGEtCCYPWCURWth5DG88bNsmY86/5kMxb1tdHQN78zYJhpo7eBzGY2yR2Habx3LG2w2GM9sOSzDwHCuTYDj2DJ8WMwmQFoMbZzcYfGw7IMEgkWMmwdhwmBgtZx4YJLbVSTDIvyFWy/keBqAtzEBbeAhpYSu2SDgHsoXNwHDGucOSbTxpQJFjuLXw8x/eePNH2W17g/OHnxnzlNXx87Mf3njjQw1uLSAgASET2MDxAY6pBLwaYFr4DzA/IKBwFIyCUTAKRigAABJES36LODBdAAAAAElFTkSuQmCC","orcid":"","institution":"Jiangxi University of Science and Technology,Nanchang","correspondingAuthor":true,"prefix":"","firstName":"Qiuming","middleName":"","lastName":"Liu","suffix":""}],"badges":[],"createdAt":"2026-04-24 06:01:49","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":true,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":true},"doi":"10.21203/rs.3.rs-9512906/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9512906/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108088847,"identity":"d02809f5-12d0-45f7-b1a3-9524f150ebca","added_by":"auto","created_at":"2026-04-29 09:03:55","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1142702,"visible":true,"origin":"","legend":"","description":"","filename":"MADASSA.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9512906/v1_covered_9f04f395-ad7f-4f3a-abf4-41e1db8007c3.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eMADA-SSA: Multi-Axis Directional Attention and Spatial-Semantic Aggregation for Robust Drone-to-Satellite Geo-Localization\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Jiangxi University of Science and Technology","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Cross-View geo-localization, Drone to satellite Matching, Strip Attention mechanism, Metric Learning, Multi-Scale Feature Fusion","lastPublishedDoi":"10.21203/rs.3.rs-9512906/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9512906/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCross-view geo-localization serves as a key technique for drone autonomous nav- igation and disaster response. Existing methods sufer from limited receptive fields, weak directional geometric modeling, and unstable metric learning under extreme perspective changes. To address these challenges, we propose a unified framework that efectively integrates three core components: a Spatial-Semantic Multi-scale Aggregation (SSMA) module, which enhances global feature rep- resentation beyond local receptive fields; a Directional Strip Attention Block (DSAB) that refines geometric structure through multi-axis gating and diago- nal remapping to bridge directional perception gaps; and a Smooth Hard-Mining Loss (SHMLoss) that stabilizes metric learning with a density-aware weighting strategy. Experiments demonstrate that our method achieves 96.01% recall@1 and 92.84% AP on the University-1652 dataset, along with consistent perfor- mance across varying altitudes on SUES-200, establishing a reliable and robust paradigm for drone-to-satellite visual localization. The code and model files are publicly available at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5281/zenodo.19702776\u003c/span\u003e\u003cspan address=\"10.5281/zenodo.19702776\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e","manuscriptTitle":"MADA-SSA: Multi-Axis Directional Attention and Spatial-Semantic Aggregation for Robust Drone-to-Satellite Geo-Localization","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-29 09:03:30","doi":"10.21203/rs.3.rs-9512906/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"55feefa4-b394-4250-865d-663d668d613d","owner":[],"postedDate":"April 29th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-29T09:03:30+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-29 09:03:30","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9512906","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9512906","identity":"rs-9512906","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}