Registered Data

[01435] Automatic generation of terrain maps using sequences of satellite images

  • Session Time & Room : 3C (Aug.23, 13:20-15:00) @E504
  • Type : Contributed Talk
  • Abstract : In this work, we propose an unsupervised methodology for analyzing temporal sequences of satellite images. Images decompose into disjoint tiles, and we embed sequences of tiles into multidimensional time series. The proposed embedding captures valuable information about the terrain and its evolution. It allows the partitioning of the ground into different types of terrain and understanding of the relationship between them. The proposed methodology shows promising results when analyzing a region of Navarre, Spain.
  • Classification : 62H35
  • Format : Talk at Waseda University
  • Author(s) :
    • Aritz Pérez (Basque Center for Applied Mathematics)
    • Carlos Echegoyen (Public University of Navarre)
    • Guzmán Santafé (Public University of Navarre)
    • Unai Pérez (Public University of Navarre)
    • María Dolores Ugarte (Public University of Navarre)