Anomaly detection based method for spatio-temporal dynamics mapping in dam mining regions

• Autor(a) principal: Gino, Vinicius L.S. [UNESP]
• Data de Publicação: 2021
• Outros Autores: Negri, Rogerio G. [UNESP], Souza, Felipe N. [UNESP]
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/239991

Remote Sensing technologies and Machine Learning methods rise as a potential combination to assemble new environmental monitoring applications. In this context, the presented work proposes a new method that exploits anomaly detection models applied to Remote Sensing imagery to identify the spatio-temporal changes over the Earth's surface. The potential of the introduced approach is shown in a study case concerning the analysis of the landscape changes using One-Class SVM and Isolation Forest methods in Landsat and Sentinel images for Brumadinho and Mariana regions, Brazil, after its re¬ cent dam collapses.