Thesis Code: 24007

Thesis type: M.Sc. in Machine Learning, Data Science, Computer Science, Mathematics, or similar

Keywords: EO data, deep learning, computer vision, aerial imagery

Requirements:

  • Knowledge of Python
  • Software development skills
  • Basic concepts of data science, concerning data analysis, processing and machine learning
  • Basic concepts of image processing, modern computer vision preferred but not required.

Description:

The escalating demand for improved spatial resolution in temperature data has propelled a heightened interest in innovative approaches that combine satellite-derived information with ancillary data sources. This thesis proposal aims to address the challenge of enhancing the spatial resolution of Sentinel-3 temperature data by integrating it with Digital Elevation Model (DEM) and land cover data.

The primary objective of this research is to develop a deep learning model capable of increasing the spatial resolution of Sentinel-3 temperature data. This will be achieved through the incorporation of DEM and land cover data to provide a more detailed and accurate representation of temperature variations across different topographic features.

The proposed methodology involves several phases. Firstly, an extensive review of existing techniques for enhancing spatial resolution in satellite-derived data will be conducted to identify the field’s most effective methodologies and eventual datasets. Subsequently, the collected Sentinel-3 temperature data will be combined with DEM and land cover data to create a comprehensive dataset.

The core of the research involves the development and training of a deep learning model. Landsat 8 satellite data, obtained concurrently with Sentinel-3 data, will be employed as supervisory information during the training phase. This supervised learning approach aims to leverage the high spatial resolution of Landsat 8 to guide the deep learning model in enhancing the spatial resolution of Sentinel-3 temperature data.

This research aims to contribute novel insights and methodologies to address the challenges associated with improving the spatial resolution of satellite-derived environmental data.

Contact: send a resume with attached the list of exams to giacomo.blanco@linksfoundation.com, luca.barco@linksfoundation.com and lorenzo.innocenti@linksfoundation.com specifying the thesis code and title.