Metodología para la evaluación de los incendios de la vegetación en la Sierra de la Macarena utilizando técnicas de análisis de imágenes basadas en objetos geográficos y árboles de decisión

Abstract

This study evaluates the severity and impact of forest fires in the Sierra de la Macarena, Meta, during the period 2018–2024, using Sentinel-2 MSI imagery and advanced geospatial analysis techniques. Through an object-based image analysis (GEOBIA) approach combined with the implementation of decision trees and sequential pattern mining, affected areas were classified, allowing not only the quantification of fire severity but also the identification of vegetation regeneration patterns over time. The results provide crucial insights into how fires have transformed the ecosystem and how it has responded over the years, offering a solid basis for post-fire planning and management in this important natural park. This comprehensive approach enabled continuous and updated monitoring of fire impacts, supporting the conservation of one of Colombia’s most biodiverse ecosystems.