[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-125741-en":3,"doc-seo-125741-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},125741,1649267921044,"Ava Thompson","https://us-avatar.wpscdn.com/avatar/1800007509477c92dfb?_k=1782875107921204101",8,"Research & Report","Machine learning and computer vision for urban drainage inspections - Publisher's version","A doctoral dissertation exploring how machine learning and computer vision can support urban drainage inspection through sewer asset management. It develops and evaluates unsupervised anomaly detection using PCA-based reconstruction, feature descriptors, dissimilarity functions, and convolutional autoencoders. It further builds convolutional neural network models for multi-label image classification, including data exploration, loss design, handling class imbalance, leave-two-inspections-out cross-validation, and image-to-pipe performance aggregation.","Machine learning and computer vision for urban drainage inspections  \nMeijer, D.W.J.  \nCitation  \nMeijer, D. W. J. (2023, November 7). Machine learning and computer vision for urban drainage inspections. Retrieved from [https://hdl.handle.net/1887/3656056](https://hdl.handle.net/1887/3656056)  \nVersion: Publisher's Version  \nLicence agreement concerning inclusion of doctoral  \nLicense:  thesis in the Institutional Repository of the University  \nof Leiden  \nDownloaded from:  [https://hdl.handle.net/1887/3656056](https://hdl.handle.net/1887/3656056)  \n[Note:](Note: To cite this publication please use the final published version)[ To cite this publication please use the final published version](Note: To cite this publication please use the final published version) (if applicable) .  \nMachine Learning and Computer Vision for Urban Drainage Inspections  \n“What do such machines really do? They increase the number of things we can do without thinking. Things we do without thinking—there’s the real danger.”  \nFrank Herbert, God Emperor of Dune  \nMachine Learning and Computer Vision for Urban Drainage Inspections  \nProefschrift  \nter verkrijging van  \nde graad van doctor aan de Universiteit Leiden, [op gezag van rector magnificus prof.dr.ir. H. Bijl](op gezag van rector magnificus prof.dr.ir. H. Bijl), volgens besluit van het college voor promoties te verdedigen op dinsdag 7 november 2023 klokke 15.00 uur  \ndoor  \nDirk Willem Johannes Meijer geboren te ‘s-Gravenhagein 1989  \nPromotores:  \ndr. A. J. Knobbe  \nprof.dr. T. H. W. Bck  \nPromotiecommissie:  \nprof.dr. M. S. K. Lew  \n[prof.dr.ir. F. J. Verbeek](prof.dr.ir. F. J. Verbeek)  \n[dr. L. Scholten](dr. L. Scholten) ([TU Delft](TU Delft))  \n[prof.dr.ir. F. H. L. R. Clemens](prof.dr.ir. F. H. L. R. Clemens) (TU Delft, NTNU Trondheim)  \ndr. K. J. Wolstencroft  \nThis work is part of the Cooperation Programme TISCA (Technology Innovation for Sewer Condition Assessment) with project number 15343, which is (partly) financed by NWO domain TTW (the domain applied and Engineering Sciences of the Netherlands Organisation for Scientific Research), the RIONED Foundation, STOWA (Foundation for Applied Water Research) and the Knowledge Program Urban Drainage (KPUD) .  \nCover design by Dirk W. J. Meijer, from data collected by Rianne A. Luimes.  \nTypeset with pdfLATEX, style based on tufte-latex.  \nPrinted by NBD Biblion Services.  \nCopyright © 2023 Dirk W. J. Meijer.  \nDedicated to the memory ofIlundi for helping me rediscover meaning.  \nContents  \n1 Introduction ........................................ 11  \n1.1 Motivation ........................................... 11  \n1.1.1 Sewer Asset Management ........................... 11  \n1.1.2 Machine Learning and Computer Vision ................ 12  \n1.1.3 Scope .......................................... 13  \n1.2 Research Questions ..................................... 13  \n1.3 Contribution ......................................... 16  \n2 Preliminaries ........................................ 18  \n2.1 Machine Learning ...................................... 18  \n2.1.1 Classification .................................... 18  \n2.1.2 Regression ...................................... 19  \n2.1.3 Overfitting, Regularization, Cross Validation ............. 20  \n2.1.4 Model Validation ................................. 23  \n2.1.5 Anomaly Detection ............................... 27  \n2.1.6 Principal Component Analysis ....................... 28  \n2.2 Digital Image Processing ................................. 29  \n2.2.1 Convolution .................................... 29  \n2.3 Convolutional Neural Networks ........................... 31  \n2.3.1 The Perceptron .................................. 31  \n2.3.2 Convolutional layers ............................... 34  \n2.3.3 Pooling layers .................................... 35  \n2.3.4 Convolutional Neural Network Design ................. 36  \n2.4 Computer Stereovision .................................. 37  \n3 Image-Based U","cbCainjnU5FnbIGp","https://ap.wps.com/l/cbCainjnU5FnbIGp","pdf",2815687,1,153,"English","en",105,"# Introduction\n## Motivation\n## Research Questions\n## Contribution\n# Preliminaries\n## Machine Learning\n## Digital Image Processing\n## Convolutional Neural Networks\n## Computer Stereovision\n# Image-Based Unsupervised Anomaly Detection\n## Framework\n## Proof of Concept\n## Application in Sewer Pipe Images\n## Convolutional Autoencoder\n## Summary\n# Convolutional Neural Network Classification\n## Introduction\n## Data Exploration\n## Methodology\n## Results\n## Discussion","[{\"question\":\"What problem does the dissertation address in urban drainage inspections?\",\"answer\":\"It addresses how machine learning and computer vision can help detect and classify conditions in urban drainage inspection images to support sewer asset management decisions.\"},{\"question\":\"How is unsupervised anomaly detection implemented?\",\"answer\":\"It combines PCA-based decomposition and partial reconstruction with feature descriptors and a dissimilarity function, then demonstrates a proof of concept and an application to sewer pipe images using a convolutional autoencoder.\"},{\"question\":\"How does the CNN classification approach evaluate model performance?\",\"answer\":\"It uses multi-label classification with a task-specific loss, manages class imbalance via oversampling, and applies leave-two-inspections-out cross-validation while aggregating performance at the pipe level.\"}]","Machine learning and computer vision for urban drainage inspections - Publisher's version | PDF",1785900972,386,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"machine-learning-and-computer-vision-for-urban-drainage-inspections-publishers-version","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/machine-learning-and-computer-vision-for-urban-drainage-inspections-publishers-version/125741/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the dissertation address in urban drainage inspections?","Question",{"text":75,"@type":76},"It addresses how machine learning and computer vision can help detect and classify conditions in urban drainage inspection images to support sewer asset management decisions.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is unsupervised anomaly detection implemented?",{"text":80,"@type":76},"It combines PCA-based decomposition and partial reconstruction with feature descriptors and a dissimilarity function, then demonstrates a proof of concept and an application to sewer pipe images using a convolutional autoencoder.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the CNN classification approach evaluate model performance?",{"text":84,"@type":76},"It uses multi-label classification with a task-specific loss, manages class imbalance via oversampling, and applies leave-two-inspections-out cross-validation while aggregating performance at the pipe level.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]