PhD student, LISTIC, USMB
Contact
E-mail: Matthieu.gallet@ univ-smb.fr
Website: https: //matthieu-gallet.github.io/
Telephone: +33(0) 4 50 09 42 09
Fax: +33(0) 4 50 09 65 59
Office: A224
Address 1: LISTIC - Polytech Annecy-Chambery, BP 80439, 74944 Annecy le Vieux Cedex, France
Thesis
Group: LISTIC
Theme: AFuTé
Topic: Machine-learning for the characterization and analysis of cryospheric dynamics by
radar satellite imagery
Summary: The texture of the cryosphere cover is variable over time as a result of energy exchanges between the snowpack, the underlying soil and the atmosphere.
between the snowpack, the underlying soil and the atmosphere. While precipitation and
and soil conditions determine the properties of a surface snow layer, its evolution over time, as well as that of the
and the different layers that are superimposed on each other to form a snowpack, depends on the
snow cover, depends on weather conditions.
The latter are assessed in collaboration with the Centre d'Étude de la Neige of Météo-France.
On a small satellite scale (a few hundred square metres), we will be interested in the analysis of
spatial variations of textures (snow in all its states and ice), while on a large scale (several hundred square kilometres), we will be interested in the presence or absence of
(several hundred square kilometres), we will be more interested in the presence or absence of snow cover or in the differentiation of
snow cover or the differentiation between dry and wet snow.
So far, the available studies on the subject are site-specific (very limited geographical areas) and
available data and even less so the potential of deep learning methods in this type of application.
of deep learning methods in this type of application.
The objective of the project is to propose so-called convolutional neural models, operating in a deep learning framework and based on
The objective of the project is to propose so-called convolutional neural models, operating in a deep learning framework and using time series of satellite images, in order to allow :
- The evaluation of the extent of cryospheric surfaces (this is called semantic segmentation);
- Automatic recognition of observable textures (classification) on cryospheric surfaces, e.g.
cryosphere, e.g.: monitoring of total snow including dry snow, thus allowing to differentiate
different snow states; - Dynamic monitoring :
- surface areas of interest (evolutions of the cryosphere feature contours in space and time) and
space and time) and - the contents of these surfaces (occupancy ratios on the same surface between the different
components/features visible on a given surface) to deduce the spatial
the spatial evolution of observable surface states.
- surface areas of interest (evolutions of the cryosphere feature contours in space and time) and
Keywords: Cryosphere, Radiative transfer, Machine learning, Radar
Supervisor: Abdourrahmane ATTO (LISTIC)
Co-supervisor: Fatima KARBOU (Météo-France), Emmanuel TROUVÉ (LISTIC)
Start of the thesis: October 2021
Doctoral School: Environmental Engineering Sciences