This repository is an archive for Matteo Luelli's Master's Thesis with the title: Emulation of High‑Spectral‑Resolution Atmospheric Radiative Transfer for Imaging Spectroscopy Earth Observation Satellite Missions. The main objective is to design AI-based emulators to obtain better performance than Look-Up Tables (LUTs) interpolation of values physically simulated with models such as MODTRAN6, while keeping fast computational speeds. This would allow such emulators to be used in hyperspectral missions such as ESA's FLEX. This task has to leverage synthetic high-resolution spectral datasets of atmospheric radiative transfer functions, generated starting from concentration and geometric parameters of the atmosphere.
The dataset employed for the first part of the thesis work is the dataset generated and used for the AMLEC challenge (Atmospheric Machine Learning Emulation Challenge).
The data can be found at this link: https://huggingface.co/datasets/isp-uv-es/rtm_emulation
The following are the hardware specifications of the main computer used to run the code, obtain results and measure runtime:
- GPU: NVIDIA GeForce RTX 4060 Ti (8 GB of VRAM)
- CPU: AMD RYZEN 9 7900 (3.70 GHz, 12 cores)
- RAM: 32 GB DDR5 with 4800 MT/s
- Storage: Crucial P310 SSD 2 NVMe pcie 4.0
Thesis is in progress.
Bibliography is in progress.