FCT - Trabalho de Conclusão de Curso

URI Permanente para esta coleção

http://200.137.215.59//handle/ri/11337

Navegar

Navegando FCT - Trabalho de Conclusão de Curso por Autor "Barros, Lucas Marques"

Agora exibindo 1 - 1 de 1
  • Nenhuma Miniatura disponível
    Item
    Aplicação de machine learning para identificação de corpos máficos-ultramáficos na porção sul da faixa Brasília, Goiás, Brasil
    (Universidade Federal de Goiás, 2024-08-02) Barros, Lucas Marques; Tabosa, Luiz Fernando Cursino; Temporim, Filipe Altoé; Temporim, Filipe Altoé; Yokoyama, Elder; Lopes, Débora Nogueira
    This research investigates the application of Machine Learning in identifying mafic-ultramafic bodies within a 200 km² area in the southern portion of the Brasília Belt, focusing specifically on the boundary between the Anicuns-Itaberaí sequence and the Anápolis-Itauçu Complex, near the municipality of Anicuns, Goiás.. The study applied the Random Forest algorithm to process and analyze geophysical and remote sensing data, aiming to enhance the geological mapping and detection of these bodies. The research integrated aerogeophysical data, including magnetometry and radiometry, as well as satellite data, in order to develop a robust geospatial dataset. The model was trained using a significant fraction of the dataset to provide a predictive map of geological occurrences. The obtained results demonstrated the high effectiveness of the model, with significant accuracy in discriminating the geological bodies of interest, which were validated by comparisons with field data and provided new insights on the distribution and characteristics of mafic-ultramafic intrusive bodies in the studied area. This study highlights the potential of using Machine Learning techniques in geology, particularly as a supporting tool for mineral prospecting, as well as providing more accurate and detailed geological maps.