Multispectral image translation from SAR data

The goal of this thesis is to investigate Deep Learning techniques for the translation of SAR data into multispectral imagery. In particular, the student will utilise a public dataset offering aligned image pairs of the Sentinel-1 (SAR) and Sentinel-2 (multispectral) satellites, such as SEN1-2 [1]. Several DL models can be explored, such as [2]-[4], and their performance will be assessed and compared.

Prerequisites: Strong Python skills, Machine Learning basic concepts, Deep Learning python framework (Pytorch, Tensorflow, etc)

Supervisor: Maria Sdraka

[1] https://mediatum.ub.tum.de/1436631
[2] https://ieeexplore.ieee.org/abstract/document/8825802
[3] https://www.sciencedirect.com/science/article/pii/S0031320321003897?casa_token=GOXb7dDI-EwAAAAA:pQPI0bBFRD5YLQ6gHkP7TiwH-RzGX9W3Xdbjz1EGG-Bp4fwQrKVVaH0l_T0uI_23Nvdq6pRi9dvA
[4] https://www.mdpi.com/2072-4292/11/17/2067