@InProceedings{CaonMeAnCaMeOl:2018:MaPaMe,
               author = "Caon, Iv{\~a} Luis and Mercante, Erivelto and Antunes, Jo{\~a}o 
                         Francisco Gon{\c{c}}alves and Cattani, Carlos Eduardo Vizzotto 
                         and Mendes, Isaque Souza and Oldoni, Lucas Volochen",
          affiliation = "{Universidade Estadual do Oeste do Paran{\'a} (UNIOESTE)} and 
                         {Universidade Estadual do Oeste do Paran{\'a} (UNIOESTE)} and 
                         {Embrapa Inform{\'a}tica Agropecu{\'a}ria} and {Universidade 
                         Estadual do Oeste do Paran{\'a} (UNIOESTE)} and {Universidade 
                         Estadual do Oeste do Paran{\'a} (UNIOESTE)}",
                title = "Mapeamento de pastagens por meio da classifica{\c{c}}{\~a}o da 
                         fus{\~a}o de imagens Landsat-8/OLI e MODIS no munic{\'{\i}}pio 
                         de S{\~a}o Gabriel do Oeste - MS",
            booktitle = "Anais...",
                 year = "2018",
               editor = "Silva, Jo{\~a}o dos Santos Vila da and Namikawa, La{\'e}rcio 
                         Massaru",
                pages = "686--694",
         organization = "Simp{\'o}sio de Geotecnologias no Pantanal 7, (GEOPANTANAL)",
            publisher = "Embrapa Inform{\'a}tica Agropecu{\'a}ria, Instituto Nacional de 
                         Pesquisas Espaciais (INPE)",
              address = "Campinas, S{\~a}o Jos{\'e} dos Campos.",
             keywords = "sensoriamento remoto, sensor orbital, processamento de imagens, 
                         minera{\c{c}}{\~a}o de dados, fus{\~a}o de imagens, 
                         classifica{\c{c}}{\~a}o de imagens, remote sensing, orbital 
                         sensor, image processing, data mining, image fusion, image 
                         classification.",
             abstract = "O sensoriamento remoto mostra-se eficiente no mapeamento de 
                         grandes {\'a}reas geogr{\'a}ficas, executado a partir de imagens 
                         orbitais. A alta resolu{\c{c}}{\~a}o espacial presente em 
                         sensores tem permitido o mapeamento detalhado da 
                         superf{\'{\i}}cie terrestre, por{\'e}m a resolu{\c{c}}{\~a}o 
                         temporal tamb{\'e}m se mostra importante, devido a constante 
                         mudan{\c{c}}a que ocorre nos ecossistemas. Desse modo os 
                         algoritmos de predi{\c{c}}{\~a}o se mostram de grande valia, uma 
                         vez que s{\~a}o capazes de unir a alta resolu{\c{c}}{\~a}o 
                         espacial de um sensor a alta resolu{\c{c}}{\~a}o temporal de 
                         outro. O objetivo deste trabalho foi realizar o mapeamento das 
                         {\'a}reas de pastagem presentes na extens{\~a}o do 
                         munic{\'{\i}}pio de S{\~a}o Gabriel do Oeste - MS, bem como 
                         avaliar o desempenho de diferentes algoritmos de 
                         classifica{\c{c}}{\~a}o em diferentes s{\'e}ries temporais, 
                         sendo uma composta apenas de imagens Landsat e outra composta de 
                         imagens geradas pelo algoritmo de predi{\c{c}}{\~a}o STARFM 
                         (Spatial and Temporal Adaptive Reflectance Fusion Model). Sendo 
                         que o algoritmo Random Forest, na s{\'e}rie temporal composta 
                         pelas imagens geradas pelo algoritmo STARFM e com a 
                         adi{\c{c}}{\~a}o de m{\'e}tricas fenol{\'o}gicas apresentou as 
                         melhores acur{\'a}cias, obtendo {\'{\i}}ndice Kappa superior a 
                         0,85 e exatid{\~a}o global superior a 92,5%. ABSTRACT: Remote 
                         sensing is efficient in the mapping of large geographic areas, 
                         executed from orbital images. The high spatial resolution present 
                         in sensors has allowed the detailed mapping of the terrestrial 
                         surface, but the temporal resolution is also important due to the 
                         constant change that occurs in the ecosystems. In this way the 
                         prediction algorithms prove to be of great value, since they are 
                         capable of joining the high spatial resolution of one sensor with 
                         high temporal resolution of another. The objective of this work 
                         was to map the pasture areas present in the extension of S{\~a}o 
                         Gabriel do Oeste - MS, as well as to evaluate the performance of 
                         different classification algorithms in different time series, one 
                         composed only of Landsat images and another composed of images 
                         generated by the STARFM (Spatial and Temporal Adaptive Reflectance 
                         Fusion Model) prediction algorithm. The Random Forest algorithm, 
                         in the time series composed of the images generated by the STARFM 
                         algorithm and the addition of phenological metrics, showed the 
                         best accuracy, obtaining a Kappa index higher than 0.85 and a 
                         global accuracy greater than 92.5%.",
  conference-location = "Jardim",
      conference-year = "20-24 out. 2018",
             language = "pt",
                  ibi = "8JMKD3MGPDW34M/46TDL68",
                  url = "http://urlib.net/ibi/8JMKD3MGPDW34M/46TDL68",
           targetfile = "p99.pdf",
                 type = "Fauna e Vegeta{\c{c}}{\~a}o",
        urlaccessdate = "08 maio 2024"
}