Simon Donike

Information

Bio

Simon Donike is a PhD researcher in Artificial Intelligence and Earth Observation at the Image and Signal Processing Group (ISP), Universitat de València. His background combines remote sensing and geospatial data science, beginning with a BSc in Geography from the University of Cologne and continuing with the Copernicus Master in Digital Earth (Erasmus Mundus), with a focus on AI for satellite imagery. Since late 2022, he has been working on an ESA-funded research project on trustworthy super-resolution of Sentinel-2 imagery, developing and analysing generative models to improve the spatial and radiometric quality of Earth Observation products for downstream scientific applications.

Current Role

He is currently a PhD researcher at the ISP in València, working within the ESA OpenSR initiative. His research focuses on generative AI for satellite image super-resolution, with particular emphasis on diffusion-based models and large-scale validation of their physical reliability in Earth Observation applications. His work is dedicated to advancing trustworthy generative modelling in remote sensing through the development of open, reproducible tools and benchmarks that bridge scientific validation and practical Earth Observation use cases.

Areas of Expertise

Simon’s expertise lies in generative AI and remote sensing, particularly diffusion- and GAN-based models for satellite image enhancement, synthesis, and gap filling. He specialises in large-scale benchmarking and validation of super-resolution methods for Sentinel-2 imagery, combining quantitative evaluation with real-world applications such as flood mapping, fire monitoring, and infrastructure analysis. These specialisations enable him to design, validate, and benchmark trustworthy generative models that enhance spatial detail, reconstruct missing information, and generate realistic synthetic data for training and testing downstream models, contributing to AI-driven Earth Observation and reproducible open-science practices within ESA Φ-lab CIN.

Vision for the Future

He is driven by a vision of a future in which generative AI enables continuous, high-fidelity Earth Observation beyond current sensor and acquisition limits. His long-term goal is to develop trustworthy models that bridge data gaps, enhance temporal and spatial coverage, and generate scientifically valid synthetic observations. By integrating validated generative modelling into EO data processing pipelines, he aims to support climate research, environmental monitoring, and rapid-response applications through open, transparent, and reliable AI systems.

Project as CIN Researcher

Coming Soon