Max Aragón Cerecedes
Second-year PhD student at Mines Paris – PSL funded by the chair Scidosol, working on physical modelling and generative AI for improving cloud retrievals and solar irradiance mapping using all-sky camera networks. Co-founder of Wematics, an Austrian startup building AI cloud cameras.
Research Projects
SkyTwin: Annotated multi-view all sky rendering
SkyTwin aims to create a comprehensive annotated dataset of multi-view all-sky images. This project will investigate the benefit of multi camera setups for cloud property retrieval.
Coming soon
SkyDiff: All sky probabilistic nowcasting using diffusion
This project focuses on developing a cloud cover nowcasting diffusion model benchmarked with a spatio temporal neural network. The generated frames are evaluated for specific tasks such as cloud cover segmentation and solar irradiance estimation.
Coming soon
SkyPix: All sky cloud genera segmentation
SkyPix is a large pixel-level dataset partitioning all sky images into semantic categories. We aim to demonstrate that finetuning a base model with few annotated key frames is sufficiently to generalise to other camera types.
Coming soon
Open-source sky image datasets for solar forecasting with deep learning: A comprehensive survey
A comprehensive review of publicly available sky image datasets used in solar forecasting applications, analyzing their characteristics, applications, and limitations.
Published in Renewable and Sustainable Energy Reviews