Docencia

• No reglada / Formació a altres
  07/05/2024
  Landscapes of Irrigation and Application of Deep Learning to Detect Qanat Water Distribution Systems
  Warsaw
  Organizador: "Near East Archaeology" Seminar. Faculty of Archaeology, University of Warsaw
  

   

  The presentation provided an introduction to landscape archaeology and deep learning for the Master’s and PhD students at the “Near East Archaeology” seminar at the Faculty of Archaeology at the University of Warsaw. The most important types of aspects of water management studies in landscape archaeology of the Near East, and the problematics of qanat studies have been discussed.

  The basic concepts of machine learning and deep learning were presented. After that, an overview of the implementation of deep learning in archaeology was demonstrated.

  The final part of the lecture consisted of a presentation of qanat detection using YOLOv8.

   

   

   

   

• No reglada / Formació a altres
  11/05/2023
  The application of satellite imagery in the landscape archaeology in the Near East. UnderTheSands: case study – Gorgan plain in Iran
  Warsaw
  Organizador: "Near East Archaeology" Seminar. Faculty of Archaeology, University of Warsaw
  

   

  The presentation provided an introduction to remote sensing for the Master’s and PhD students at the «Near East Archaeology» seminar at the Faculty of Archaeology at the University of Warsaw. The most important types of applications of satellite imagery in the landscape studies were discussed, such as landscape destruction, looting, site detection and mapping of architectural structures and water management.  

    

  The practical application of remote sensing was then demonstrated in the examples of the study of the UnderTheSands project, which aims to locate and reconstruct the irrigation network using remote sensing and machine / deep learning methods in the Near East. The presentation showed an application of the methods developed in UnderTheSands on the example of Gorgan Plain (Iran), one of the critical areas of study on settlement patterns and qanat irrigation in the Near East (Hopper 2017).  

    

  The presented methods allowed us to explore the potential of applying multitemporal satellite data processed in the Google Earth Engine (Sentinel 2, LANDSAT 5) and the application of recently declassified Hexagon imagery and TanDEM-X digital model in the landscape studies (© DLR 2022, DEM_HYDR3723).