%0 Conference Proceedings
%4 sid.inpe.br/marte2/2017/10.23.19.34.40
%2 sid.inpe.br/marte2/2017/10.23.19.34.41
%@isbn 978-85-17-00088-1
%F 59233
%T Otimização de classificação supervisionada da cobertura do solo em São Leopoldo (RS) por meio de seleção de conjuntos de dados mínimos
%D 2017
%A Ribeiro, Bárbara Maria Giaccom,
%A Centeno, Jorge Antonio Silva,
%A Mendes, Carlos André Bullhões,
%@electronicmailaddress bgiaccom@gmail.com
%E Gherardi, Douglas Francisco Marcolino,
%E Aragão, Luiz Eduardo Oliveira e Cruz de,
%B Simpósio Brasileiro de Sensoriamento Remoto, 18 (SBSR)
%C Santos
%8 28-31 maio 2017
%I Instituto Nacional de Pesquisas Espaciais (INPE)
%J São José dos Campos
%P 1321-1328
%S Anais
%1 Instituto Nacional de Pesquisas Espaciais (INPE)
%X Recent developments in geotechnologies have provide resources to propose innovative strategies for urban and environmental management, including remote sensing data and computational resources for processing them, which together, can generate high-quality map products and valuable databases. For the purpose of mapping the Earth''s surface, digital image processing and classification enables information extraction through recognition of patterns and objects related to features of interest. The practical use of large volumes of orbital data implies, however, some costs, for example, the computational cost, which is generally high, and is required for data processing and classification. In many cases, one faces a classification problem resulting from non-increase of results accuracy as the number of bands used (and therefore the amount of information available) increases. One possible solution lies in selecting a subset of features with more discriminating power among the available bands. The aim of this study is to evaluate and compare the performance of land cover classification using a parametric classifier (Maximum Likelihood) using different sets of input data (i.e., number of spectral bands), extracted from two Landsat 8 images (dry × rainy seasons), city of São Leopoldo, Rio Grande do Sul, Brazil. The data sets are defined based on the calculation of the transformed divergence. Finally, the results are analyzed statistically to assess the quality of the classifications.
%9 Processamento de imagens
%@language pt
%3 59233.pdf