Logotipo do repositório
 

Publicação:
Land use and land cover classification using hyperspectral imagery: Evaluating the performance of spectral angle mapper, support vector machine and random forest

dc.contributor.authorChristovam, L. E. [UNESP]
dc.contributor.authorPessoa, G. G. [UNESP]
dc.contributor.authorShimabukuro, M. H. [UNESP]
dc.contributor.authorGalo, M. L.B.T. [UNESP]
dc.contributor.institutionUniversidade Estadual Paulista (UNESP)
dc.date.accessioned2022-04-29T08:28:05Z
dc.date.available2022-04-29T08:28:05Z
dc.date.issued2019-06-04
dc.description.abstractLand Use and Land Cover (LULC) information is an important data source for modeling environmental variables, so it is essential to develop high quality LULC maps. The hundreds of continuous spectral bands gathered with hyperspectral sensors provide high spectral detail and consequently confirm hyperspectral remote sensing as an appropriate option for many LULC applications. Despite increased spectral detail, issues like high dimensionality, huge volume of data and redundant information, mean that hyperspectral image classification is a complex task. It is therefore essential to develop classification approaches that deals with these issues. Since classification results are directly dependent on the dataset used, it is fundamental to compare and validate the classification approaches in public datasets. With this in mind, aiming to provide a baseline, four classification models in the relatively new hyperspectral HyRANK dataset were evaluated. The classification models were defined with three well-known classification algorithms: Spectral Angle Mapper (SAM), Support Vector Machine (SVM) and Random Forest (RF). A classification model with SAM and another with RF were defined with the 176 surface reflectance bands. A dimensionality reduction with principal component analysis was carried out and a classification model with SVM and another with RF were defined using 14 principal components as features. The results show that SVM and RF algorithms outperformed by far the SAM in terms of accuracy, and that the RF is slightly better than the SVM in this respect. It is also possible to see from the results that the use of principal components as features provided an improvement in the accuracy of the RF and an improvement of 28% in the time spent fitting the classification model.en
dc.description.affiliationGraduate Program in Cartographic Sciences São Paulo State University
dc.description.affiliationDept. of Mathematics and Computer Sciences São Paulo State University
dc.description.affiliationDept. of Cartography São Paulo State University
dc.description.affiliationUnespGraduate Program in Cartographic Sciences São Paulo State University
dc.description.affiliationUnespDept. of Mathematics and Computer Sciences São Paulo State University
dc.description.affiliationUnespDept. of Cartography São Paulo State University
dc.format.extent1841-1847
dc.identifierhttp://dx.doi.org/10.5194/isprs-archives-XLII-2-W13-1841-2019
dc.identifier.citationInternational Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives, v. 42, n. 2/W13, p. 1841-1847, 2019.
dc.identifier.doi10.5194/isprs-archives-XLII-2-W13-1841-2019
dc.identifier.issn1682-1750
dc.identifier.scopus2-s2.0-85067524416
dc.identifier.urihttp://hdl.handle.net/11449/228693
dc.language.isoeng
dc.relation.ispartofInternational Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives
dc.sourceScopus
dc.subjectHyperspectral Classification
dc.subjectHyRANK dataset
dc.subjectLULC
dc.subjectRF
dc.subjectSAM
dc.subjectSVM
dc.titleLand use and land cover classification using hyperspectral imagery: Evaluating the performance of spectral angle mapper, support vector machine and random foresten
dc.typeTrabalho apresentado em evento
dspace.entity.typePublication
unesp.departmentCartografia - FCTpt

Arquivos