openEO Platform takes off at the ESA Living Planet Symposium 2022

Michele Claus
openEO Platform
Published in
5 min readJul 4, 2022

--

The Living Planet Symposium (LPS) 2022 in Bonn from 23.05.2022 to 27.05.2022 has been a valuable milestone for openEO Platform where we could reach an uncountable number of Earth Observation experts and researchers through a great number of presentations, demos, and trainings.

If you don’t know what openEO Platform is yet, take a look at our previous blog post or visit our main website!

The openEO Platform numbers at LPS

Let’s demonstrate the success story of openEO Platform at the conference based on some numbers

  • 1 user consultation event
  • 1 poster
  • 3 oral presentations
  • 5 classroom trainings
  • 6 demos

The OpenEO Platform Team kicked off on Monday the LPS with the demo Parcel delineation from Sentinel-2 data using deep learning¹. It showed how extensible openEO Platform is: by leveraging User Defined Functions (UDF) in addition to the standard processes, it is possible to use the deep learning library TensorFlow directly in the cloud.

Tuesday was focused on the first interactive training, which described multiple approaches of Analysis-Ready Data processing², where multiple approaches were described for both optical and radar ARD processing. Subsequently, an additional classroom presented the possibility of openEO Platform: Publishing your openEO workflow as a service³. You have a fancy algorithm that cost you tons of hours to develop? openEO offers you the possibility to transform it into an operational service! The day was finalized by Demonstrating the Web Editor and R-Client⁴, where two openEO clients were introduced: firstly, the user-friendly Web Editor, which is an essential tool for people who are not used to programming but like to use interfaces such as model builders. Secondly the R-client, a full-fledged R-package with multiple integrated tools to interact with an openEO back-end.

From left to right: Peter Zellner (EURAC Research) and Matthias Mohr (Münster University) demonstrating the Web Editor and the R client at LPS

Wednesday was THE day for openEO Platform! It began with an oral presentation entitled Continental scale crop type mapping using openEO Platform⁵, which showed the scalability of the platform at its best: multiple cloud infrastructures are employed behind the scenes (Terrascope, Sentinel Hub, Creodias) to generate a European crop type map covering EU27 (4M km2) — 400 MGRS tiles. Being able to use multiple back-ends is a key factor since it allows to share the processing and not being blocked if a single infrastructure is down. Subsequently, another presentation with the title Cloud platform federation through openEO, new concepts and implementations in the ESA openEO Platform⁶ introduced the overall project, highlighted the advantages and also challenges of the openEO Platform federation.

From left to right: Jeroen Dries (VITO) and Alexander Jacob (EURAC Research) during the Big EO data — Platforms and Applications & Products session.

On the same day, two classroom trainings were completed. The first training covered the topic Using openEO to support deep learning use cases⁷, a specific machine learning use case that focuses on the inference step, using a trained model. The second one called openEO Platform: Introduction to the Web editor and R-Client⁸ represented a sequel to the demo given about the web editor and the R-Client and gave participants the opportunity for hands-on training in these two openEO clients.

The day ended with the User Consultation Event⁹, a special meeting for openEO Platform research grant applicants, early adopters, developers and the project management team. During the event, the research grant applicants presented their interesting use cases implemented using the openEO Platform and at the end, a jury awarded the best ideas.

After this fully packed day, the conference was still going on and we still had two days to share our vision.
On Thursday, the oral presentation entitled The R-spatial package ecosystem and openEO for analysing Earth Observation¹⁰ data showed the specific potential of the recently updated openEO R-client. Another classroom training named Advanced data processing with openEO Platform¹¹ held on that day, where some advanced openEO workflows for snow monitoring has been shown and explained. The day ended with another demo about analysis ready data processing.

On Friday, the final day of the conference, we were still on-site sharing our insights about our forest canopy cover use case, firstly with a demo named Large scale fractional canopy cover mapping with openEO platform¹² and later with a poster¹³ session about it. This use case involved multiple cloud infrastructures as well, with a PlanetScope very high-resolution data preprocessing running at VITO and the machine learning model training and testing running at EODC.

Looking back at this wonderful and exciting experience, we are proud to have shared our work with the EO community coming from several scientific disciplines. The positive feedbacks and numerous discussions around this project ensured us there is a great potential for future work benefitting from the various data and processing capabilities openEO Platform is offering.

openEO Platform is funded by the European Space Agency (ESA).

References

  1. Parcel delineation from Sentinel-2 data using deep learning
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3994&ses=21466#anker_session_21466
    GitHub: https://github.com/openEOPlatform/parcel-delineation
  2. openEO Platform: Analysis-Ready Data processing
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3996&ses=21495#anker_session_21495
    GitHub: https://github.com/openEOPlatform/LPS_presentations/tree/main/demos/EODC_ARD_processing
  3. Publishing your openEO workflow as a service
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3995&ses=21520#anker_session_21520
    GitHub: https://github.com/openEOPlatform/LPS_presentations/tree/main/classrooms/VITO_workflow_as_service
  4. openEO Platform: Demonstrating the Web Editor and R-Client
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3995&ses=21479#anker_session_21479
  5. Continental scale crop type mapping using openEO platform https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3996&ses=20738#
  6. Cloud platform federation through openEO, new concepts and implementations in the ESA openEO Platform
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3996&ses=20740#
  7. Using openEO to support deep learning use cases
    https://express.converia.de/index.php?page_id=18446&v=List&do=15&day=3996&ses=21524#anker_session_21524
    GitHub: https://github.com/openEOPlatform/LPS_presentations/blob/main/classrooms/VITO_deep_learning/notebooks.md
  8. openEO Platform: Introduction to the Web editor and R-Client
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3996&ses=21527#anker_session_21527
    GitHub: https://github.com/openEOPlatform/LPS_presentations/tree/main/classrooms/WWU_Web_Editor_Rclient
  9. User consultation event
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3996&ses=21150#anker_session_21150
  10. The R-spatial package ecosystem and openEO for analysing Earth Observation data
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3997&ses=20828#anker_session_20828
  11. Advanced data processing with openEO Platform
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3997&ses=21532#anker_session_21532
    GitHub: https://github.com/openEOPlatform/LPS_presentations/tree/main/classrooms/EURAC_advanced_processing
  12. Large scale fractional canopy cover mapping with openEO platform
    https://express.converia.de/frontend/index.php?page_id=18446&v=List&do=15&day=3998&ses=21516#anker_session_21516
  13. Cloud-based Forest Change and Fractional Canopy Cover Retrieval using the federated openEO Platform infrastructure
    https://express.converia.de/frontend/index.php?page_id=22770&additions_conferenceschedule_action=detail&additions_conferenceschedule_controller=paperList&pid=65836&hash=ce99f77efffedfc00be84495277a6fba4a83a14da15baa4a174b93ed1f858254

--

--

Michele Claus
openEO Platform

Researcher at EURAC — Institute for Earth Observation