Ensuring reliable water level measurement for flooding: A redundancy-based approach with pressure transducer and computer vision

Registro completo de metadados
MetadadosDescriçãoIdioma
Autor(es): dc.contributorUniversidade de São Paulo (USP)-
Autor(es): dc.contributorUniversidade Estadual Paulista (UNESP)-
Autor(es): dc.creatorMatos, Saulo Neves-
Autor(es): dc.creatorRocha, Arthur Lima Marques-
Autor(es): dc.creatorDomingues Filho, Gabriel Montagni-
Autor(es): dc.creatorRanieri, Caetano Mazzoni-
Autor(es): dc.creatorGarcia, Rodrigo Dutra-
Autor(es): dc.creatorFaria, Ana Clara de Oliveira-
Autor(es): dc.creatorMedina, Maria Mercedes Gamboa-
Autor(es): dc.creatorUeyama, J.-
Data de aceite: dc.date.accessioned2025-08-21T20:21:55Z-
Data de disponibilização: dc.date.available2025-08-21T20:21:55Z-
Data de envio: dc.date.issued2025-04-29-
Data de envio: dc.date.issued2024-12-31-
Fonte completa do material: dc.identifierhttp://dx.doi.org/10.1177/01423312241285952-
Fonte completa do material: dc.identifierhttps://hdl.handle.net/11449/302900-
Fonte: dc.identifier.urihttp://educapes.capes.gov.br/handle/11449/302900-
Descrição: dc.descriptionFluid level measurement is essential in many fields, including industrial and civil sectors, especially for urban flood detection, where there is a high risk of mortality and economic losses. However, although contact-based methods that employ pressure transducers can achieve a high degree of precision, they are susceptible to damage from direct contact with the fluid. This study adopts a redundancy-based approach that combines pressure transducer measurements with computer vision to provide enhanced reliability and reduce the risk of sensor failures. Our approach entails training a deep-learning model that uses pressure sensor data to mitigate this potential risk of damage and avoid the need for manually annotating sets of images. The results show that the pressure transducer has high accuracy, with a mean absolute error (MAE) of 1.21 cm, and that the computer vision model which is trained on pressure sensor data, achieves a comparable MAE of 6.67 cm. This approach also makes the system more robust and includes a dependable backup measurement method in case the primary sensor fails. Furthermore, the model trained on the sensor data led to results that were very similar to those trained directly on ground-truth data.-
Descrição: dc.descriptionInstitute of Mathematical and Computer Sciences University of São Paulo (USP), SP-
Descrição: dc.descriptionSão Carlos School of Engineering University of São Paulo (USP), SP-
Descrição: dc.descriptionSão Carlos Institute of Physics University of São Paulo (USP), SP-
Descrição: dc.descriptionInstitute of Geosciences and Exact Sciences São Paulo State University (UNESP), SP-
Descrição: dc.descriptionInstitute of Geosciences and Exact Sciences São Paulo State University (UNESP), SP-
Idioma: dc.languageen-
Relação: dc.relationTransactions of the Institute of Measurement and Control-
???dc.source???: dc.sourceScopus-
Palavras-chave: dc.subjectComputer vision-
Palavras-chave: dc.subjectdeep learning-
Palavras-chave: dc.subjectflood prediction-
Palavras-chave: dc.subjectpressure transducer-
Palavras-chave: dc.subjectwater level-
Título: dc.titleEnsuring reliable water level measurement for flooding: A redundancy-based approach with pressure transducer and computer vision-
Tipo de arquivo: dc.typelivro digital-
Aparece nas coleções:Repositório Institucional - Unesp

Não existem arquivos associados a este item.