Optimized Multi-Reference Detection for Edge Devices in Flood Monitoring: A Model Quantization Approach

Optimized Multi-Reference Detection for Edge Devices in Flood Monitoring: A Model Quantization Approach | 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 Optimized Multi-Reference Detection for Edge Devices in Flood Monitoring: A Model Quantization Approach Zhihong Yang, Jinhuan Wang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7019876/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 Multi-reference detection for flood depth estimation is a challenging task demanding high accuracy, real-time performance, and adaptability across diverse scenarios. This study proposes a model quantization strategy leveraging knowledge distillation to construct a high-precision, real-time detection model tailored for edge devices, such as street-view cameras. By obtaining a teacher model with high accuracy and low inference speed, and a student model with lower accuracy but higher inference speed, knowledge distillation is applied to enhance the student model's detection accuracy by 10.64% in Mask's mAP50, while preserving its inference speed. Compared to YOLO11m, the teacher model achieves an 8% increase in Mask's mAP50. The distilled student model reduces parameters and GFLOPs by 34.98% and 27.43%, respectively, relative to YOLO11m, and by 33.23% and 30.9% compared to the teacher model. This framework offers an effective solution for edge devices in flood monitoring, balancing accuracy and real-time performance. Project details are available via the following link: https://github.com/muhan-yy/FloodDepth_LightWeight. Edge Devices Multi-Reference Detection YOLO11 Knowledge Distillation Full Text Additional Declarations No competing interests reported. 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-7019876","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":497780593,"identity":"ad5af783-0b6f-4258-9ea9-226c8c53fa58","order_by":0,"name":"Zhihong Yang","email":"","orcid":"","institution":"The University of Auckland","correspondingAuthor":false,"prefix":"","firstName":"Zhihong","middleName":"","lastName":"Yang","suffix":""},{"id":497780595,"identity":"7f3175f5-27a8-4919-8e6c-3001759ef426","order_by":1,"name":"Jinhuan Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA60lEQVRIiWNgGAWjYLACyQYJOX72xoYDHwxs7IjVYmEs2XP44MEZBWnJxGlhbKhI3HAjLfkwz4dDjA2EVBscP3v4heUOicSZPWcMDtsYHGBmYD98dANeLWfy0iwkz0gY97P3GBzOMbjDx8CTlnYDnxazAzlmBpJtErJgW3IMnjEzSPCY4ddy/g1YC+OGGzkGhy0MDjM2ENRyI8f4AVCLItD7CYcZiNFif+ONGQNQCyiQDxzsMUhLZiPkF8n+HOPPkm11oKhs/vDjj40dP/vhY3i1AAGbtAQKl4ByEGD++IEIVaNgFIyCUTCCAQCaRFRQItazQgAAAABJRU5ErkJggg==","orcid":"","institution":"YanShan University","correspondingAuthor":true,"prefix":"","firstName":"Jinhuan","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2025-07-01 11:38:28","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7019876/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7019876/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102748627,"identity":"b10de796-65b6-4501-8d22-832b3565cc4c","added_by":"auto","created_at":"2026-02-16 09:11:19","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":687585,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7019876/v1_covered_a844481d-1ec8-45b9-9674-bbc3811eb4ab.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Optimized Multi-Reference Detection for Edge Devices in Flood Monitoring: A Model Quantization Approach","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","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":"Edge Devices, Multi-Reference Detection, YOLO11, Knowledge Distillation","lastPublishedDoi":"10.21203/rs.3.rs-7019876/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7019876/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMulti-reference detection for flood depth estimation is a challenging task demanding high accuracy, real-time performance, and adaptability across diverse scenarios. This study proposes a model quantization strategy leveraging knowledge distillation to construct a high-precision, real-time detection model tailored for edge devices, such as street-view cameras. By obtaining a teacher model with high accuracy and low inference speed, and a student model with lower accuracy but higher inference speed, knowledge distillation is applied to enhance the student model's detection accuracy by 10.64% in Mask's mAP50, while preserving its inference speed. Compared to YOLO11m, the teacher model achieves an 8% increase in Mask's mAP50. The distilled student model reduces parameters and GFLOPs by 34.98% and 27.43%, respectively, relative to YOLO11m, and by 33.23% and 30.9% compared to the teacher model. This framework offers an effective solution for edge devices in flood monitoring, balancing accuracy and real-time performance. Project details are available via the following link: https://github.com/muhan-yy/FloodDepth_LightWeight.\u003c/p\u003e","manuscriptTitle":"Optimized Multi-Reference Detection for Edge Devices in Flood Monitoring: A Model Quantization Approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-14 11:28:53","doi":"10.21203/rs.3.rs-7019876/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":"e64e9da6-d384-4b7b-80ca-99e1fac045be","owner":[],"postedDate":"August 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-02-13T22:38:43+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-14 11:28:53","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7019876","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7019876","identity":"rs-7019876","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}