@InProceedings{MoreiraAlmCruFurSoa:2015:MoCoAp,
               author = "Moreira, Mayne Assun{\c{c}}{\~a}o and Almeida, Paula Maria Moura 
                         de and Cruz, Carla Bernadete Madureira and Furtado, Luiz Felipe de 
                         Almeida and Soares, Mario Luiz Gomes",
          affiliation = "{} and {} and {} and {Instituto Nacional de Pesquisas Espaciais 
                         (INPE)}",
                title = "Modelagem do conhecimento aplicada {\`a} detec{\c{c}}{\~a}o de 
                         mudan{\c{c}}as em ambiente costeiro",
            booktitle = "Anais...",
                 year = "2015",
               editor = "Gherardi, Douglas Francisco Marcolino and Arag{\~a}o, Luiz 
                         Eduardo Oliveira e Cruz de",
                pages = "2023--2030",
         organization = "Simp{\'o}sio Brasileiro de Sensoriamento Remoto, 17. (SBSR)",
            publisher = "Instituto Nacional de Pesquisas Espaciais (INPE)",
              address = "S{\~a}o Jos{\'e} dos Campos",
             abstract = "The landscape is constantly changing, be it natural or 
                         anthropogenic character. Coastal environments are naturally more 
                         dynamic than the inner portions of the continent, and lately has 
                         been suffering with landscape changes by anthropogenic action. 
                         Aiming at monitoring of these environments, the study of the 
                         landscape changes has always been the target of numerous studies 
                         of remote sensing. At the same time, the techniques used for such 
                         analyzes has been constant improvement, however, a major challenge 
                         is still analyzing large time series in such dynamic environment 
                         as coastal areas. In this context, the present work was developed 
                         in the pursuit of optimizing change detection techniques, without 
                         losing the quality of the product generated. Using a historical 
                         series of nine TM/Landsat 5 images, with 30 km resolution, 
                         covering the period 1984-2006, and object-based images analysis, a 
                         multiresolution segmentation of bands 3 and 4 each scene was done. 
                         The classification of areas of change was made in two levels of 
                         segmentation using mainly descriptors NDVI 
                         (m{\'{\i}}n/m{\'a}x_NDVI and amp_NDVI). The result showed that 
                         the optimization technique and the descriptors used were very 
                         efficient for the separability of the classes not change and 
                         change, with very good global accuracy (0.81) and Kappa index 
                         (0.76) at 1: 150,000 scale, validated based reference points 
                         collected in the field.",
  conference-location = "Jo{\~a}o Pessoa",
      conference-year = "25-29 abr. 2015",
                 isbn = "978-85-17-0076-8",
                label = "401",
             language = "pt",
         organisation = "Instituto Nacional de Pesquisas Espaciais (INPE)",
                  ibi = "8JMKD3MGP6W34M/3JM49HL",
                  url = "http://urlib.net/ibi/8JMKD3MGP6W34M/3JM49HL",
           targetfile = "p0401.pdf",
                 type = "An{\'a}lise de s{\'e}ries de tempo de imagens de sat{\'e}lite",
        urlaccessdate = "27 abr. 2024"
}