Earth Observation & Weather Data Federation with AI Embeddings

Informazioni relative al progetto

Embed2Scale

ID dell’accordo di sovvenzione: 101131841

DOI

10.3030/101131841

Data della firma CE 11 Dicembre 2023

Data di avvio 1 Gennaio 2024

Data di completamento 31 Dicembre 2026

Finanziato da

• Digital, Industry and Space

Costo totale

€ 2 618 047,50

Contributo UE

€ 2 618 047,50

Coordinato da

MARTEL INNOVATE BV
Netherlands

CORDIS fornisce collegamenti ai risultati finali pubblici e alle pubblicazioni dei progetti ORIZZONTE.

I link ai risultati e alle pubblicazioni dei progetti del 7° PQ, così come i link ad alcuni tipi di risultati specifici come dataset e software, sono recuperati dinamicamente da .OpenAIRE .

Risultati finali

Documents, reports (8)

Use Case Implementation Report

This deliverable will describe the implementation features of the Use Cases of the first development phase, also considering the requirements.

Analysis of ARD data and APIs

Report on current state of the ARD and associated data processing APIs

Data access for AI-compressors

Documentation, complete source code and configuration files. Pull Requests on OpenEO and OpenDataCube git repositories.

Reference Code for AI-compressors

Python code as sample reference of the AI-compressor methodologies for Earth observation

Dissemination and Communication Strategy and Plan

This deliverable provides the dissemination and communication foundation, including the strategic guidance to direct Embed2Scale community building activities for broad and durable impact. It will be led by Task 5.1.

SSL4EO-S12-downstream benchmark dataset

Provides an open-source benchmark dataset to assess AI-compressor methodologies

AI-compressor containerization CICD pipeline

Documentation, complete source code and configuration files

Report on State-of-the-Art in AI-compressors

Review on AI-compressors in the light of the Embed2Scale application pillars Embed2{Train,Infer,Transfer,Find}

Other (1)

Benchmarking environment for AI-compressors

Benchmarking environment to evaluate AI-compressor performance available to run an Earth observation community data challenge

Pubblicazioni

Conference proceedings (2)

Earth Observation Applications through Neural Embedding Compression from Foundation Models

Autori: Carlos Gomes, Thomas Brunschwiler
Pubblicato in: 2024
Editore: Copernicus GmbH
DOI: 10.5194/egusphere-egu24-19460

Geospatial Foundation Models for Efficient Retrieval of Remote Sensing Images

Autori: Brunschwiler, T., Blumenstiel, B., Moor, V., and Kienzler, R.
Pubblicato in: 2024
Editore: EGU General Assembly 2024
DOI: 10.5194/egusphere-egu24-18300

Other (3)

Scalable Efficient Compression in Large-Scale Earth Observation

Autori: Erik Scheurer, Jiangtao Wang, Rocco Sedona, Stefano Maurogiovanni, Benedikt Blumenstiel, Johannes Jakubik, Paolo Fraccaro, Thomas Brunschwiler, Stefan Kesselheim, Gabriele Cavallaro
Pubblicato in: 2025
Editore: Copernicus GmbH
DOI: 10.5194/EGUSPHERE-EGU25-19016

"Neural Embedding Compression for Earth Observation Data – an Ablation Study"

Autori: Amelie Koch, Isabelle Wittmann, Carlos Gomez, Rikard Vinge, Michael Marszalek, Conrad Albrecht, Thomas Brunschwiler
Pubblicato in: 2025
Editore: Copernicus GmbH
DOI: 10.5194/EGUSPHERE-EGU25-17172

Earth Observation embeddings at the test: A novel benchmark to evaluate (neural) compression for satellite imagery

Autori: Rikard Vinge, Michael L Marszalek, Jannik Schneider, Conrad M Albrecht
Pubblicato in: 2025
Editore: Copernicus GmbH
DOI: 10.5194/EGUSPHERE-EGU25-16756