@article{LázaroGómez:147498,
      author        = "Lázaro Gómez, Javier and Rodrigues Mimbrero, Marcos",
      title         = "{IDENTIFICACIÓN DE PARCELAS AGRÍCOLAS ABANDONADAS EN LA
                       COMARCA DEL MATARRAÑA-MATARRANYA MEDIANTE TELEDETECCIÓN Y
                       TÉCNICAS DE APRENDIZAJE ARTIFICIAL}",
      year          = "2023",
      note          = "Classification and regression models, together with
                       environmental and landscape studies and research through
                       remote sensing and GIS, are among the main lines of
                       research at present due to their capacity to address
                       territorial studies at regional scales. In Mediterranean
                       areas, their application for the analysis of the process of
                       agricultural land abandonment is contributing to the
                       monitoring and better understanding of this phenomenon. The
                       Matarraña/Matarranya region has undergone numerous changes
                       in agricultural land and vegetation in recent decades,
                       especially due to demographic, environmental, landscape,
                       economic and political factors. The present Final Degree
                       Project aims to create a classification model of the main
                       land covers involved in the dynamics of abandonment,
                       differentiating between crops, abandoned plots and natural
                       vegetation, over the period of time covering the years
                       1984-2015 in the Matarraña region. The collection and
                       analysis of the data has been through Remote Sensing and
                       GIS, using Landsat satellite images, and subsequent
                       processing through the Landtrendr and R environment.
                       Results have been obtained related to trends in spectral
                       response in terms of brightness and humidity (components of
                       the Tasseled Cap transformation), thus being able to
                       predict the evolution and identification of the various
                       plots in the study area with an accuracy of 80% accuracy in
                       the classification by Random Forest. This type of study
                       allows us to deepen in the multitemporal analysis of land
                       cover, a component of great importance to understand the
                       territorial dynamics for its management and planning.",
}