DATA MINING TECHNIQUES APPLIED TO ALOS-2/PALSAR-2 SATELLITE IMAGERY FOR LAND USE AND LAND COVER CLASSIFICATION

This paper proposes a workflow for the classification of synthetic aperture radar (SAR) images obtained by the ALOS-2/PALSAR-2 satellite, aiming at the land use and land cover mapping. The study area is located in the western portion of Federal District of Brazil. The presented approach combines multiresolution segmentation, object attributes, and iterative machine learning procedures. A set of 397 attributes was generated based on the amplitude images, HH and HV polarizations. These attributes were processed in the WEKA 3.8 software using the J48 decision tree, Random Forest and Multilayer Perceptron Artificial Neural Network classifiers. Classification results attained Kappa indices higher than 0.70, especially the Multilayer Perceptron Artificial Neural Network algorithm (Kappa = 0.87). This workflow demands low time processing and has potential to be reproduced for other study sites or SAR images obtained at different wavelengths.