Representatives from academia, research, industry, policy, etc., of the milestone event organized by ESA “2022 Living Planet Symposium” on the 23rd -27th of May 2022 [World Conference Center Bonn, Germany], will have the opportunity to plunge into the EIFFEL world through an elaborative paper-poster.
This symposium will focus on how Earth Observation contributes to science and society, and how disruptive technologies and actors are changing the traditional Earth Observation landscape, which is also creating new opportunities for public and private sector interactions.
EIFFEL will participate with the paper-poster entitled “Deep Learning Based Sentinel-2 Super Resolution Methods for Supporting Climate Change Adaptation & Mitigation Applications”.
Europe’s efforts towards the implementation of the Paris Agreement for Climate Change (CC) are substantial. The Earth Observation data from the Sentinel program provide the required information to support CC adaptation and mitigation policies. Despite the abundance of data in GEOSS and their undeniable value, their uptake in the context of CC applications is often limited due to inherent spatial resolution constraints. To address this challenge, the EIFFEL Horizon2020 project will create tools to enhance the spatial resolution of Earth Observation data available in GEOSS to address the needs of five different CC adaptation and mitigation applications.
The way forward:
EIFFEL has developed a large training dataset with a global distribution of samples to compare the performance of deep residual network architectures and their hyperparameters and to select the most efficient which has global applicability for Sentinel 2 Level 2A data without retraining. Additionally, the super-resolved images were used in a CC-related application to demonstrate their value in such applications. As a future work, this tool will provide super-resolved Sentinel-2 data as input to five different CC related applications which will be tested on real life scenarios. The actual enhancement of the application results will be the evidence of the value of Sentinel-2 DL super-resolution techniques in supporting CC mitigation and adaptation applications.
More info here
Find the full paper-poster here