%0 Conference Proceedings
%4 dpi.inpe.br/sbsr@80/2008/11.16.21.05
%2 dpi.inpe.br/sbsr@80/2008/11.16.21.05.08
%@isbn 978-85-17-00044-7
%T Estratificação espacial utilizando árvores de decisão para estimativa da área de culturas agrícolas de verão com imagens de pré-plantio
%D 2009
%A Arcoverde, Gustavo Felipe Balué,
%A Epiphanio, José Carlos Neves,
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@electronicmailaddress gustavo@dsr.inpe.br
%@electronicmailaddress epiphanio@dsr.inpe.br
%E Epiphanio, José Carlos Neves,
%E Galvão, Lênio Soares,
%B Simpósio Brasileiro de Sensoriamento Remoto, 14 (SBSR)
%C Natal
%8 25-30 abr. 2009
%I Instituto Nacional de Pesquisas Espaciais (INPE)
%J São José dos Campos
%P 75-82
%S Anais
%1 Instituto Nacional de Pesquisas Espaciais (INPE)
%K crop area estimation, remote sensing, spatial stratification, pattern recognition, tree decision.
%X The improvement of the estimative of area of summer crops are important for many users. In general, stratifications are applied for these estimates. In this study we evaluate the process of spatial stratification through tree decisions. The stratifications usually to improve the statistical variances associating when coupled to field verification. Pattern recognition of summer crop on pre-planting TM/Landsat images using tree decision was used to generate spatial stratifications. For this purpose, multi-features were extracted from the satellite images in Barretos municipality, São Paulo State, in Brazil. The tree decision was generate setting the minimum instances by leaves in order to reduce the size of the trees. As results, the variance and area estimates using stratifications has not a significant improvement in relation to the area estimates without stratification, based on a Monte Carlo method. This study showed that in small areas for crop areas to be estimated, the size of the defined stratum (a single stratum) should be nearly the half of whole area to be estimated. Otherwise, the tree decision had a good performance to recognize patterns.
%9 Agricultura
%@language pt
%3 75-82.pdf