Comparison of Machine Learning Algorithms for Mapping the Phytophysiognomies of the Brazilian Cerrado

Luciano T. de OliveiraThomaz C. de A. OliveiraLuiz M. T. de CarvalhoWilian Soares LacerdaSamuel R. de S. CamposAdriana Z. Martinhago

This present work describes the classification of the Phytophysiognomies present in the Brazilian Cerrado biome through the means Artificial Intelligence; data from remote sensing images and other sources served as input for these algorithms to generate the vegetation maps. The data acquired was of many types so that it fully described the various Phytophysiognomies present in biome and served as training data for the machine learning algorithms. Various statistical and neuro-computation based algorithms were used for pattern recognition in the data so that we could build a good generalization model for the biome. A vegetation map was successfully generated with each algorithm. Finally a comparison among these algorithms was made so that we could find the best algorithm that fitted the problem of mapping this biome.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: