Latest Tweets

Explainable AI: application to trustworthy super-resolution (OpenSR) 

UNIVERSITY OF OXFORD (GB)

Summary

The project aims to bring robust, accountable, and scalable multi-spectral super-resolution techniques to the Earth Observation (EO) community for the ubiquitous L2 and L3 pre-processing of the Sentinel-2 (S2) revisits archive. Super-resolution (SR) is a nascent technology and the roadmap to maturity will require insights from many disciplines. Super-resolution is not just about image generation, but also degradation: how much is lost in pixelation. To shift the public perception on the safety of SR-S2 products, the project will provide uncertainty and quality metrics along with the SR products; establish and disseminate best practices through new methods and tools that will be open to everyone.
The project will push the boundaries of excellence science and technological development of SR in remote sensing, by creating a set of tools, platform, and guidelines:
  • Tools of state-of-the-art SR, Explainable AI (xAI), saliency and information metrics.
  • An open WebGIS platform that goes beyond standard solutions by working directly on Analysis Ready Data stored in a datacube. The platform will bring SR S2 data at the fingertips of users, allow for interactive data exploration, analysis and combination of SR S2 and other data, and even on-the-fly execution of xAI models on S2 data. It will provide access to the data from a number of use cases illustrating different real-world problems and showcasing how applications benefit from the combined action of SR and xAI, and their limits.
  • A set of guidelines and best practices that summarize an accountable and reproducible SR pipeline for successful applications.

Information

Domain
AI4EO
Prime contractor
UNIVERSITY OF OXFORD (GB)
Subcontractors
  • BROCKMANN CONSULT GMBH (DE)
  • UNIV VALENCIA (ES)