{"id":{"repo_id":"brazil-uerj","oai_identifier":"oai:pantheon.ufrj.br:11422/6316"},"canonical_url":"https://search.dev.ndltd.org/etd/brazil-uerj/oai:pantheon.ufrj.br:11422/6316","repository":{"repo_id":"brazil-uerj","name":"Brazil UERJ","base_url":"https://pantheon.ufrj.br/oai/request"},"display":{"title":"Deep CNN and MLP-based vision systems for algae detection in automatic inspection of underwater pipelines","abstract":"Artificial neural networks, such as the multilayer perceptron (MLP), have been increasingly employed in various applications. Recently, deep neural networks, specially convolutional neural networks (CNN), have received considerable attention due to their ability to extract and represent high-level abstractions in data sets. This work describes a vision inspection system based on deep learning and computer vision algorithms for detection of algae in underwater pipelines. The proposed algorithm comprises a CNN or a MLP network, followed by a post-processing stage operating in spatial and temporal domains, employing clustering of neighboring detection positions and a region interception framebuffer. The performances of MLP, employing different descriptors, and CNN classifiers are compared in real-world scenarios. It is shown that the post-processing stage considerably decreases the number of false positives, resulting in an accuracy rate of 99.39%.","abstract_html":"Artificial neural networks, such as the multilayer perceptron (MLP), have been increasingly employed in various applications. Recently, deep neural networks, specially convolutional neural networks (CNN), have received considerable attention due to their ability to extract and represent high-level abstractions in data sets. This work describes a vision inspection system based on deep learning and computer vision algorithms for detection of algae in underwater pipelines. The proposed algorithm comprises a CNN or a MLP network, followed by a post-processing stage operating in spatial and temporal domains, employing clustering of neighboring detection positions and a region interception framebuffer. The performances of MLP, employing different descriptors, and CNN classifiers are compared in real-world scenarios. It is shown that the post-processing stage considerably decreases the number of false positives, resulting in an accuracy rate of 99.39%.","abstract_has_math":false,"creators":["Medina Castañeda, Edgar Eduardo"],"institution":"Universidade Federal do Rio de Janeiro","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Petraglia, Mariane Rembold"],"committee_chairs":[],"committee_members":[],"year":2017,"date_issued":"2017-10","date_published":"2017-10","updated_at":"2026-07-24T01:16:21Z","subjects":["Engenharia elétrica","Redes neurais convolucionais","Visão computacional"],"languages":["eng"],"rights":["Acesso Aberto"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/11422/6316","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Petraglia, Mariane Rembold"]},{"key":"dc:creator","label":"Author","values":["Medina Castañeda, Edgar Eduardo"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2019-01-31T14:54:50Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2026-05-16T03:03:43Z"]},{"key":"dc:date.issued","label":"Date","values":["2017-10"]},{"key":"dc:publisher","label":"Institution","values":["Universidade Federal do Rio de Janeiro"]},{"key":"dc:publisher.department","label":"Dc Publisher Department","values":["Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia"]},{"key":"dc:type","label":"Dc Type","values":["Dissertação"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Engenharia elétrica","Redes neurais convolucionais","Visão computacional"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Acesso Aberto"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["http://hdl.handle.net/11422/6316"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Artificial neural networks, such as the multilayer perceptron (MLP), have been increasingly employed in various applications. Recently, deep neural networks, specially convolutional neural networks (CNN), have received considerable attention due to their ability to extract and represent high-level abstractions in data sets. This work describes a vision inspection system based on deep learning and computer vision algorithms for detection of algae in underwater pipelines. The proposed algorithm comprises a CNN or a MLP network, followed by a post-processing stage operating in spatial and temporal domains, employing clustering of neighboring detection positions and a region interception framebuffer. The performances of MLP, employing different descriptors, and CNN classifiers are compared in real-world scenarios. It is shown that the post-processing stage considerably decreases the number of false positives, resulting in an accuracy rate of 99.39%."]},{"key":"dc:title","label":"Title","values":["Deep CNN and MLP-based vision systems for algae detection in automatic inspection of underwater pipelines"]}]}],"canonical_facts":{"dc:contributor.advisor":["Petraglia, Mariane Rembold"],"dc:creator":["Medina Castañeda, Edgar Eduardo"],"dc:date.accessioned":["2019-01-31T14:54:50Z"],"dc:date.available":["2026-05-16T03:03:43Z"],"dc:date.issued":["2017-10"],"dc:description.abstract":["Artificial neural networks, such as the multilayer perceptron (MLP), have been increasingly employed in various applications. Recently, deep neural networks, specially convolutional neural networks (CNN), have received considerable attention due to their ability to extract and represent high-level abstractions in data sets. This work describes a vision inspection system based on deep learning and computer vision algorithms for detection of algae in underwater pipelines. The proposed algorithm comprises a CNN or a MLP network, followed by a post-processing stage operating in spatial and temporal domains, employing clustering of neighboring detection positions and a region interception framebuffer. The performances of MLP, employing different descriptors, and CNN classifiers are compared in real-world scenarios. It is shown that the post-processing stage considerably decreases the number of false positives, resulting in an accuracy rate of 99.39%."],"dc:identifier.uri":["http://hdl.handle.net/11422/6316"],"dc:language":["eng"],"dc:publisher":["Universidade Federal do Rio de Janeiro"],"dc:publisher.department":["Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia"],"dc:rights":["Acesso Aberto"],"dc:subject":["Engenharia elétrica","Redes neurais convolucionais","Visão computacional"],"dc:title":["Deep CNN and MLP-based vision systems for algae detection in automatic inspection of underwater pipelines"],"dc:type":["Dissertação"]},"updated_at":"2026-07-24T01:16:21Z"}