@InProceedings{BendiniFMMTSHV:2021:ExDeCo,
               author = "Bendini, Hugo do Nascimento and Fonseca, Leila Maria Garcia and 
                         Maretto, Raian Vargas and Matosak, Bruno Menini and Taquary, 
                         Evandro Carrijo and Sim{\~o}es, Philipe Souza and Haidar, Ricardo 
                         and Valeriano, Dalton de Morisson",
          affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto 
                         Nacional de Pesquisas Espaciais (INPE)} and {University of Twente} 
                         and {Instituto Nacional de Pesquisas Espaciais (INPE)} and 
                         {Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto 
                         Nacional de Pesquisas Espaciais (INPE)} and {UniversidadeFederal 
                         do Tocantins (UFTO)} and {Instituto Nacional de Pesquisas 
                         Espaciais (INPE)}",
                title = "Exploring a deep convolutional neural network and geobia for 
                         automatic recognition of brazilian palm swamps (veredas) using 
                         Santinel-2 optical data",
            booktitle = "Proceedings...",
                 year = "2021",
         organization = "International Geoscience and Remote Sensing Symposium (IGARSS)",
            publisher = "IEEE",
              address = "Breussels",
             keywords = "Cerrado, Semantic Segmentation, Peatlands, Remote Sensing, Digital 
                         Processing Image.",
             abstract = "The Brazilian Palm Swamps (Veredas) are a vegetation physiognomy 
                         of the Cerrado biome. It has a critical importance for 
                         biodiversity and also for groundwater sources conservation. With 
                         the irrigated agriculture intensification, it īs been 
                         significantly impacted. Mapping this physiognomy is important to 
                         delimit this vegetation type to provide subsides for public policy 
                         and monitoring programs. Pixel-based methods do not succeed, since 
                         the spatial context is important for this physiognomy. 
                         Object-based methods are a great potential on this sense. Deep 
                         Learning methods, particularly the convolutional neural networks 
                         (CNN), are increasing considerably as a solution for these 
                         challenges. We applied both methods in two regions of the Cerrado 
                         and evaluated the model transferability. The results are 
                         promising, with training model overall accuracies higher than 90% 
                         for both methods. The CNN performed better when transferred a 
                         different region. We discussed some advantages and limitations, 
                         and pointed out to improvements that can still be done.",
  conference-location = "Online",
      conference-year = "12-16 July",
           targetfile = "bendini_2021.pdf",
        urlaccessdate = "09 maio 2024"
}