A Machine Learning Strategy Based on Kittler’s Taxonomy to Detect Anomalies and Recognize Contexts Applied to Monitor Water Bodies in Environments

Página do item simplificado
dc.contributor.authorDias, Maurício Araújo [UNESP]
dc.contributor.authorMarinho, Giovanna Carreira [UNESP]
dc.contributor.authorNegri, Rogério Galante [UNESP]
dc.contributor.authorCasaca, Wallace [UNESP]
dc.contributor.authorMuñoz, Ignácio Bravo
dc.contributor.authorEler, Danilo Medeiros [UNESP]
dc.contributor.institutionUniversidade Estadual Paulista (UNESP)
dc.contributor.institutionUniversity of Alcalá (UAH)
dc.date.accessioned2023-03-01T20:42:11Z
dc.date.available2023-03-01T20:42:11Z
dc.date.issued2022-05-01
dc.description.abstractEnvironmental monitoring, such as analyses of water bodies to detect anomalies, is recog-nized worldwide as a task necessary to reduce the impacts arising from pollution. However, the large number of data available to be analyzed in different contexts, such as in an image time series acquired by satellites, still pose challenges for the detection of anomalies, even when using comput-ers. This study describes a machine learning strategy based on Kittler’s taxonomy to detect anomalies related to water pollution in an image time series. We propose this strategy to monitor environ-ments, detecting unexpected conditions that may occur (i.e., detecting outliers), and identifying those outliers in accordance with Kittler’s taxonomy (i.e., detecting anomalies). According to our strategy, contextual and non-contextual image classifications were semi-automatically compared to find any divergence that indicates the presence of one type of anomaly defined by the taxonomy. In our strategy, models built to classify a single image were used to classify an image time series due to domain adaptation. The results 99.07%, 99.99%, 99.07%, and 99.53% were achieved by our strat-egy, respectively, for accuracy, precision, recall, and F-measure. These results suggest that our strategy allows computers to recognize contexts and enhances their capabilities to solve contextualized problems. Therefore, our strategy can be used to guide computational systems to make different decisions to solve a problem in response to each context. The proposed strategy is relevant for im-proving machine learning, as its use allows computers to have a more organized learning process. Our strategy is presented with respect to its applicability to help monitor environmental disasters. A minor limitation was found in the results caused by the use of domain adaptation. This type of limitation is fairly common when using domain adaptation, and therefore has no significance. Even so, future work should investigate other techniques for transfer learning.en
dc.description.affiliationDepartment of Mathematics and Computer Science Faculty of Sciences and Technology São Paulo State University (UNESP), Campus Presidente Prudente
dc.description.affiliationDepartment of Environmental Engineering Sciences and Technology Institute São Paulo State University (UNESP), Campus São José dos Campos
dc.description.affiliationDepartment of Energy Engineering São Paulo State University (UNESP), Campus Rosana
dc.description.affiliationPolytechnic School University of Alcalá (UAH)
dc.description.affiliationUnespDepartment of Mathematics and Computer Science Faculty of Sciences and Technology São Paulo State University (UNESP), Campus Presidente Prudente
dc.description.affiliationUnespDepartment of Environmental Engineering Sciences and Technology Institute São Paulo State University (UNESP), Campus São José dos Campos
dc.description.affiliationUnespDepartment of Energy Engineering São Paulo State University (UNESP), Campus Rosana
dc.description.sponsorshipCoordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
dc.description.sponsorshipFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.description.sponsorshipIdFAPESP: 2016/24185-8
dc.description.sponsorshipIdFAPESP: 2020/06477-7
dc.description.sponsorshipIdFAPESP: 2021/01305-6
dc.identifierhttp://dx.doi.org/10.3390/rs14092222
dc.identifier.citationRemote Sensing, v. 14, n. 9, 2022.
dc.identifier.doi10.3390/rs14092222
dc.identifier.issn2072-4292
dc.identifier.scopus2-s2.0-85130079980
dc.identifier.urihttp://hdl.handle.net/11449/240991
dc.language.isoeng
dc.relation.ispartofRemote Sensing
dc.sourceScopus
dc.subjectanomaly detection
dc.subjectKittler’s taxonomy
dc.subjectmachine learning
dc.subjectpattern recognition
dc.subjectremote sensing
dc.subjecttime series
dc.titleA Machine Learning Strategy Based on Kittler’s Taxonomy to Detect Anomalies and Recognize Contexts Applied to Monitor Water Bodies in Environmentsen
dc.typeArtigo
unesp.departmentMatemática e Computação - FCTpt

Arquivos

Coleções

Presidente Prudente - FCT - Faculdade de Ciências e Tecnologia
Artigos - Turismo - Rosana