Meteorological conditions are determinant for the agricultural production; in particular, rainfall may be cited as the most important because having direct relation with water balance. To estimate agricultural production, agrometeorological models based on the cultures behavior under meteorological conditions, have been used. Since it is difficult to obtain the required data to these models, rainfall estimation techniques using meteorological satellites images from spectral channels have been used. The objective of the present work is to apply the Optimum-Path Forest pattern classifier to the agrometeorological research field in order to correlate the available information from GOES-12 satellite infrared spectral channel images, to the reflectivity data obtained by the IPMET/UNESP radar located at Bauru, aiming to develop a model for precipitation occurrence identification. In the experiments we compared four classification algorithms: Artificial Neural Networks (ANN), k-Nearest Neighbors (k-NN), Support vector Machines (SVM) and optimum-Path Forest (OPF). this last one shows the best results in terms of accuracy rate and running time.
Supervised Classifiers; Optimum-Path Forest; GOES; Precipitation Occurrence Identification
