Artificial neural network-based method to classify sedimentary rocks

Rodrigo Marques de Figueiredo, Maurício Roberto Veronez, Francisco Manoel Wohnrath Tognoli, Marcio Rosa da Silva, Luiz Gonzaga Jr., Lucas Silveira Kupssinskü, Fabiane Bordin, Diego Brum, Caroline Lessio Cazarin

Abstract: In a geological study, an important step is to determine the type of sedimentary rock or its grain size. Such a determination requires accurate analysis in the field or in a laboratory. As the size of the study area grows, this activity can be time consuming and error prone because the number of specialists working under rigid criteria also increases. This paper proposes a novel methodology to classify grain size using unique wavelength reflectance data and artificial neural networks. The results indicate that the proposed method can be reliably used in the field.