@InProceedings{FranciscoAlme:2013:MiDaAn,
               author = "Francisco, Cristiane Nunes and Almeida, Cl{\'a}udia Maria de",
          affiliation = "{} and {Instituto Nacional de Pesquisas Espaciais (INPE)}",
                title = "Minera{\c{c}}{\~a}o de dados e an{\'a}lise de imagens baseada 
                         em objeto aplicadas ao mapeamento de cobertura da terra",
            booktitle = "Anais...",
                 year = "2013",
               editor = "Epiphanio, Jos{\'e} Carlos Neves and Galv{\~a}o, L{\^e}nio 
                         Soares",
                pages = "2282--2289",
         organization = "Simp{\'o}sio Brasileiro de Sensoriamento Remoto, 16. (SBSR)",
            publisher = "Instituto Nacional de Pesquisas Espaciais (INPE)",
              address = "S{\~a}o Jos{\'e} dos Campos",
             abstract = "This article is committed to evaluate the performance of a 
                         semantic network generated by data mining for the classification 
                         of land cover using GEographic Object-Based Image Analysis 
                         (GEOBIA) in a tropical mountainous area. The study area 
                         corresponds to the Nova Friburgo County, with an extension of 933 
                         kmē, located in the region of the Fluminense Ridge, presenting 
                         thus a steep mountainous relief. Based on a visual interpretation 
                         of the images, eight land cover classes were defined: rockies, 
                         forest, grasslands, sparse grasslands, burn scars, reforestation, 
                         shadow, and urban areas. The dataset used for data mining was 
                         composed by 130 attributes and by 225 training samples accounting 
                         for all land cover classes. The algorithm C4.5, implemented in the 
                         software Weka 3.6.4, was employed for the data mining procedure. 
                         The following attributes were selected by C4.5: NDVI, fourth 
                         principal component, second angular moment, homogeneity, entropy, 
                         and slope. The obtained global accuracy was 88%, and the Kappa 
                         index reached 0.81. Only the class urban areas presented omission 
                         errors greater than 50%, being confused in some cases with sparse 
                         grasslands, forest, and burn scars. In view of the obtained value 
                         for the Kappa index, we can state that the classification 
                         presented an excellent accuracy according to a rating scale 
                         specially elaborated for such index.",
  conference-location = "Foz do Igua{\c{c}}u",
      conference-year = "13-18 abr. 2013",
                 isbn = "{978-85-17-00066-9 (Internet)} and {978-85-17-00065-2 (DVD)}",
                label = "68",
             language = "pt",
         organisation = "Instituto Nacional de Pesquisas Espaciais (INPE)",
                  ibi = "3ERPFQRTRW34M/3E7G6HN",
                  url = "http://urlib.net/ibi/3ERPFQRTRW34M/3E7G6HN",
           targetfile = "p0068.pdf",
                 type = "Classifica{\c{c}}{\~a}o e Minera{\c{c}}{\~a}o de Dados",
        urlaccessdate = "09 maio 2024"
}