An R implementation for Bayesian networks applied to spatial data
Márcio Pupin Mello, Bernardo Friedrich Theodor Rudorff, Marcos Adami, Daniel Alves Aguiar · Procedia Environmental Sciences · 2011
This work aimed to develop an R algorithm for land use classification based on the relationships among the land use and variables associated to its occurrence on remote sensing images. The algorithm was tested for soybean crop identification in the Brazilian Soy Moratorium context. Probability functions were modeled based on the number of pixels within discrete intervals. The result was encouraging with overall classification accuracy greater than 80%, indicating that the method is promising also to be applied for other land use classifications. The R algorithm is available at http://www.dsr.inpe.br/∼mello.