SIGN IN

See the results of the third round of the Sentinel Hub Custom Script Contest here. Visit our GitHub repository to explore all custom scripts.



Sentinel Hub Custom Script Contest

Remote sensing is making its way into more and more applications and fields. Tools like the Sentinel Hub EO Browser with its custom scripting allow all people to browse, analyse and visualise satellite data in almost real-time.

Together with our partners, the Copernicus EU Earth Observation programme, the European Space Agency and AI4EO consortium we are looking for new and innovative ideas and scripts. Therefore, we have organized a series of the Sentinel Hub Custom Script Contests starting with the first one in Spring 2019 and following with the second one in Winter 2019/2020. In April and May we have organized a special edition of the Contest - the Covid-19 Custom Script Contest. It was aiming to help the world facing huge challenges getting back to 'normal' after the coronavirus Covid-19 pandemic. The success of the series motivated us to continue with the third edition of the Contest in Summer 2020.

Results

The third round of the Contest introduced a new challenge (optional), with dedicated prizes. To qualify for it, besides the already known requirements, scripts had to be based on a machine learning workflow developed using eo-learn - a collection of open source Python packages to seamlessly access and automate the processing of spatio-temporal image sequences acquired by any satellite fleet. You can find an example of such script here. Ideally participants contributed to the code-base during the contest.

A selection of visualisations from the second round of the Contest

A selection of visualisations created with the custom scripts submitted to the second round of the Sentinel Hub Custom Script Contest. Check also one of our blog posts for more.

Categories

We wanted the Contest to be completely open. So, all the submitted scripts competed under one common category no matter for which data source (Sentinels, Landsat, etc.) the scripts were used for, or if the script was eo-learn-based or not. Many scripts could run in different categories anyway, so it was just fair to compare them to all scripts handed in.

How to Create a Script?

If you need some help and instructions on how to get started with custom scripting for the EO Browser we suggest you read:

If you are new to the custom scripting, we would suggest that you check out our friendly Custom Script Tutorial, which will guide you through the first steps and help you create simple custom scripts.

The Forest Cut Temporal Detection script - the second runner in the first Contest under the Agriculture and Forestry category - applied to the Sentinel-2 image of Madagascar forest from 2018 and compared with a raw image from 2017. Visualization in Sentinel Playground Temporal.

Attractive Prizes

We awarded the best three custom scripts as follows:

  • First prize: 1,000 EUR and 1-year Enterprise Sentinel Hub account (worth 5,000 EUR)
  • Second prize: 300 EUR and 1-year Individual Commercial Sentinel Hub account (worth 1,000 EUR)
  • Third prize: 1-year Individual Commercial Sentinel Hub account (worth 1,000 EUR)

In addition to above, special prizes were awarded to:

  • the best eo-learn-based script: 700 EUR
  • the best custom script submitted by a high-schooler or a high-school team: 300 EUR + a box of goodies large enough for all team members - There wasn't any submissions received from a high-scooler or a high-school team.

Among all the early bird submissions we awarded:

  • the best custom script with 500 EUR
  • the best eo-learn-based script with 300 EUR - There wasn't any submissions including eo-learn-based scripts received on time.

The scripts winning the early bird awards also competed in the overall competition at the end of the Contest. All winners will also get an opportunity to highlight their winning scripts in the Sentinel Hub blog.

How to Participate?

  • All scripts must run within the Sentinel Hub EO Browser or with multi-temporal instance of Sentinel Playground.
  • All scripts had to be handed in as text files until 30th November 2020 (23:59 CET).
  • The early bird submission needed to be handed in before 20th September 2020 (23:59 CET).
  • Together with the script, you had to hand in a proper description on what your script can be used for and explain how it is achieving that. You could include the description in Jupyter Notebook or in a separate file.
  • By handing in a script, you confirmed to be the sole author(s) of the script and to have the necessary rights to do so.
  • Every single participant or group could hand in up to three scripts.
  • All entries must be licensed under the Creative Commons BY license, so they can be added to the custom scripts repository and be used freely by all.

Besides, the eo-learn-based scripts also needed to comply with the following:

  • The process has to cover a machine learning technique and it has to use eo-learn. The result of the process has to be an actual custom script.
  • You had to submit Jupyter Notebook or Python code (as an open-source).
The Forest Hurricane script - the third best SAR script in the first Contest - applied to the Sentinel-1 image of Chojnice, Poland in August 2017, showing the area before and after the hurricane. Visualization in EO Browser.

Additional Requirements

Besides the general description of the script the following information was required:

  • Scripts applicability
  • False detection problems (if any)
  • Detailed description of how the script works
  • Representative cases / images with details
  • Visualization in EO Browser or with multi-temporal instance of Sentinel Playground (a link to a working script)
  • References
  • Credits - In case your script is based on other people's work you had to make sure to include the necessary citations and links to relevant papers.

By participating in the Contest, the participants confirmed that they have read and agreed with the Terms and Conditions of the Sentinel Hub Custom Script Contest as well as with the Sentinel Hub's Terms of Service.

Timelapse (2019) of Lake Balaton with Sentinel-2 and Sentinel-3 imagery side by side showing the Ulyssys Water Quality Viewer - the winning script in the second round of the Contest. Click here for more about the script.

Our Jury

Annamaria Luongo

Remote sensing consultant. Her main interests are Earth surface deformation caused by volcanoes and earthquakes, deforestation and natural hazards. She also has a keen interest in climate change and pollution.

Harel Dan
GIS and remote sensing at HaMaarag, Israel's National Nature Assessment Program. His special interests are wildfire damages, long-term vegetation trends, landcover and landcover change mapping, and the encroaching of buildings and infrastructure on open landscapes.

Hayley Evers-King
Hayley Evers-King is a marine EO scientist with a focus on Sentinel-3 and ocean colour. Hayley has worked on a wide variety of topics including water quality, ocean heat flux, and carbon pools. She is passionate about growing the use of EO data and currently supports EUMETSAT on user support and training activities for their Copernicus missions. She believes science communication is part of every scientists job and has been involved in many activities promoting EO science, from the creation of ocean MOOCs, youtube videos, and open source python tutorials, to music compositions, poetry evenings, and cooking demonstrations.

Mark Parrington
Mark Parrington is a senior scientist at ECMWF where he works with the Copernicus Atmosphere Monitoring Service development team. His scientific interests are in using observations from satellites and aircraft to understand wildfires and how their emissions impact global atmospheric composition and air quality.

Miha Kadunc
Chief Technology Officer and Co-founder of Sinergise

Stef Lhermitte
Remote sensing scientist with specific interest in the use of multi-source remote sensing and land surface modelling to assess the cryosphere, atmosphere and ecosystem dynamics. Since 2016 he is assistant professor of Geoscience & Remote Sensing at the TUDelft in the Netherlands.

Thomas E L Smith
Dr Thomas Smith is Assistant Professor of Environmental Geography in the Department of Geography & Environment at the London School of Economics. Dr Smith is particularly interested in complex interactions between agricultural practices, land degradation, wildfires emissions and their associated impacts.

Evaluation Criteria

  • Originality
  • Scientific relevance
  • Global applicability (scripts, which can run on different parts of the world rank better)
  • Business/Commercial potential
  • Visual appeal

Burned areas after fires

Burned areas after fires - multi-temporal script showing places, where NBR changed significantly before and after fire. View script in Sentinel Playground.

 

FAQ

  • When and how will I be informed about results?
    • We will announce the winners of the Contest on this web page latest on 30th November 2020. However, everyone will also be informed by e-mail within 3 working days after the announcement.
  • Can I base the script on a publicly available research paper?
    • Yes, of course. We believe there is a lot of useful research available in the EO domain and it should be used to all extent. You should however appropriately credit the work in the accompanying document. 
  • Do I need to subscribe to Sentinel Hub to participate in the Contest?
    • No. EO Browser is a free tool. In case you need some advanced features (e.g. time-lapse, comparison, etc.), you can create a free account. If you would actually require Sentinel Hub web services to fine-tune your script, you can use a 30-days free trial option or simply send us an e-mail and we will ensure you can do your work.
  • My script stops working after it becomes too long.
    • There is indeed a technical limit on number of chars, which can be passed over HTTP. But no worries, there is a simple way to solve it.
  • I am more interested in sharing knowledge than a monetary prize. Should I still participate?
    • Of course, this is the main objective of this Contest. In case you do not want to receive the prize, you can name a philanthropy organisation of your choice and we will donate the amount to them.
  • Can I submit scripts, which just look nice?
    • Main goal of this action is to share knowledge. However, art is important as well and we need to be aware of the beauty of our Planet, so go ahead!
  • If I do not win the prize, will my script still be added to the Custom Script repository?
    • We plan to add all scripts, which seem sound, to the repository. We will also be publishing a stream of short blog posts about all interesting scripts.
  • I have submitted a script but now I want to correct it. What to do?
    • No problem. Simply submit a new version of the script and state very clearly at the beginning that this submission is replacing a previous one.
  • Where can I get more information?
    • Ask a question on our Forum and we will respond.

See also Terms and Conditions of the Sentinel Hub Custom Script Contest for more details.

 

Results of the First Custom Script Contest

All the scripts that were submitted to the Contest in Spring 2019 were classified in one of the categories below and added to the GitHub (see results below for direct links). The winners of the contest were announced during the Living Planet Symposium in Milan on May 15th 2019 and informed by e-mail the day after.

The authors of the best scripts won attractive prizes, and the author of the overall best script got an invitation to visit the European Space Research Institute (ESRIN), home of ESA’s center for Earth Observation, to discuss his script with experts in the relevant fields.

The winning scripts were highlighted in Why join the next Sentinel Hub Custom Script Contest blog post and in two guest blog posts, Mapping deforestation with Sentinel Hub by Nicolas Karasiak, and Sentinel Hub supporting IBISA in the Crop Micro-Insurance Field by Jean-Baptiste Pleynet.

Best Overall Script

Detecting deep moist convection with Sentinel-2 - Stavros Dafis, Greece

Agriculture and Forestry

  1. Anomaly Detection - Jean-Baptiste Pleynet, Luxemburg
  2. Forest cut temporal detection - Nicolas Karasiak, France
  3. Pseudo Forest Canopy Density (pseudo-FCD) - Antonio Carlón Paredes, Spain

Marine Environment

  1. White-water detection - Mohor Gartner, Slovenia
  2. Ocean plastic detector prototype - Bence Mélykúti, DPhil (Oxf), Germany

Snow and Glaciers

  1. Monthly snow report - Nicolas Karasiak, France
  2. Snow Cover Change Detection - Karl Chastko, Canada

Disaster Management and Prevention

  1. Detecting deep moist convection with Sentinel-2 - Stavros Dafis, Greece
  2. Seldom and regular water surface detection - Mohor Gartner, Slovenia
  3. Fire boundary script - Adzanil Rachmadhi Putra, Fatwa Ramdani, Indonesia

Air Quality and Urban Planning

  1. Green City - Carlos Bentes, Estonia
  2. City Highlights - Thales Sehn Körting, Brazil
  3. Sentinel-1 Urban Areas - Monja Šebela, Slovenia

Other Scripts

  1. Land surface temperature (LST) mapping - Mohor Gartner, Slovenia
  2. Monthly composite - Nicolas Karasiak, France
  3. Barren Soil Script - Monja Šebela, Slovenia

Best SAR Script

  1. Seldom and regular water surface detection - Mohor Gartner, Slovenia
  2. Sentinel-1 Urban Areas - Monja Šebela, Slovenia
  3. Forest Hurricane - Kamil Onoszko, Poland

 

Results of the Second Custom Script Contest

The second Contest started on October 15th 2019 and ended on January 31st 2020. All entries were judged by a jury of experts on functionality and usefulness, as well as possible commercial potential.

The winning scripts were highlighted in Step into the Beautiful World of Custom Scripts blog post.

The Winning Scripts

1. Ulyssys Water Quality Viewer by András Zlinszky, Hungary
2. SAR-Ice: a Sea Ice RGB composite by Martin Raspaud, Sweden and Mikhail Itkin, Norway
3. Satellite Derived Bathymetry Mapping (SDBM) by Mohor Gartner, Slovenia

UWQV script - Tsimlyansk Reservoir, Russia

Tsimlyansk Reservoir, Russia. The Ulyssys Water Quality Viewer applied to the Sentinel-2 image, acquired on September 5th, 2019. The script is dynamically visualizing the chlorophyll and sediment conditions of water bodies on both Sentinel-2 and Sentinel-3 images See it in EO Browser.

All received scripts are available on our Github repository:

4. Tonemapped natural color script by Gregory Ivanov, Russia
5. Aquatic plants and algae custom script detector (APA-script) by Anna Péliová, Carla Garcia-Lozano and Josep Sitjar, Spain
6. Water In Wetlands Index (WIW) - Sentinel-2 Version by WILLM Loïc, LEFEBVRE GaÎtan, DAVRANCHE AurÈlie, CAMPAGNA Julie, REDMOND Lauren, MERLE ClÈment, GUELMAMI Anis and POULIN Brigitte, France
7. Water In Wetlands Index - Landsat 8 Version by WILLM Loïc, LEFEBVRE GaÎtan, DAVRANCHE AurÈlie, CAMPAGNA Julie, REDMOND Lauren, MERLE ClÈment, GUELMAMI Anis and POULIN Brigitte, France
8. Se2WaQ by Nuno Sidónio Andrade Pereira, Portugal
9. Soil Moisture Estimation script by Narayana Rao Bhogapurapu, India
10. Land use visualization for Sentinel 2 using Linear Discriminant Analysis by Marta Elvira, Roberto Calvo and Javier Becerra, Spain
11. OLCI Natural Colours with Sigmoid by Jérôme LOUIS, France
12. TOA Ratio B09-B8A ColorMap Blue-Red & Natural Colours by Jérôme LOUIS, France
13. Selective Enhancement based on Indices by Sérgio Augusto Jardim Volkmer, Brazil
14. Water Bodies' Mapping - WBM by Mohor Gartner, Slovenia
15. Tracking Radar Vegetation Index (agriculture development) change by Valters Zeizis, Latvia
16. Flood Mapping With Sentinel-1 by Maryam Salehi, Iran
17. Index visualization by Martin Javorka, Slovakia
18. Homage to Mondrian by Matevž Pintar, Slovenia
19. NDVI on L2A Vegetation and natural Colours by Jérôme LOUIS, France
20. Radar Vegetation Index for Sentinel-1 (RVI4S1) SAR data by Dipankar Mandal, India
21. Radar vegetation index code for dual polarimetric - RVI script by Subhadip Dey, India
22. SAR for deforestation detection by Antonio Carlón Paredes, Spain
23. Agricultural Crop Monitoring from Space by Maryam Salehi, Iran

 

Results of the COVID-19 Custom Script Contest

European Space Agency (ESA), in coordination with the European Commission, has launched a special edition of the Custom Script Contest, focused on the support of space assets during the COVID-19 crisis, managed by Euro Data Cube group. Following a similar format, but further to looking for new algorithms, we were in the quest for ideas on how satellite data could help monitor and mitigate the situation after the coronavirus Covid-19 pandemic. During the Contest (from 6th of April to 31st of May 2020) we have received many interesting ideas and proposals from remote sensing experts, machine learning scientists and the interested public.


The Best Overall Idea


Monthly Winning Ideas - April 2020


Monthly Winning Ideas - May 2020


Weekly Winning Ideas

The best ideas were presented to the European Commission in the context of the dossier ESA is preparing on space as a support to Covid-19 situation in the EU. Read more about the winning scripts of RACE Upscaling Competition in one of our blog posts.


 

Results of the Third Custom Script Contest

The third Contest started on August 4th 2020 and ended on November 30st 2020. All entries were judged by a jury of experts on functionality and usefulness, as well as possible commercial potential.

The winning scripts will be highlighted on our Sentinel Hub Blog in coming weeks. So, stay tuned!

The Winning Scripts

1. Reactiv Script by Thomas di Martino, Elise Colin-Koeniguer, Regis Guinvarc’h, and Laetitia Thirion-Lefevre, France
2. PUCK - Perceptually-Uniform Color Map Kit by Keenan Ganz, USA
3. Landslide detection for rapid Mapping Using Sentinel 2 by Alexander Ariza and Norma Davila, Germany

The best eo-learn-based script:

The early bird submission prize went to Thomas di Martino, Elise Colin-Koeniguer, Regis Guinvarc’h, and Laetitia Thirion-Lefevre for their Reactiv Script.

The received scripts are available on our Github repository:

4. Landsat 8 Clouds Segmentation Script by Ingrid Grenet and Houssem Farhat, France
5. Oil Slicks and Red Tide Monitoring by TIZNEGAR Startup Co, Iran
6. Mix Leaf Area Index (MixLAI) by Hamid Rahimi, Iran
7. DeepGNDVI Script by Roshni Biswas and Anurag Saha Roy, USA
8. Vegetation and Land Monitoring with Cloud Mask by TIZNEGAR Startup Co, Iran
9. DeepBAIS2 by Roshni Biswas and Anurag Saha Roy, USA
10. Active Fire Detection S2 by TIZNEGAR Startup Co, Iran
11. High Accuracy Active Fire Point Detection by TIZNEGAR Startup Co, Iran
12. Detection of Evapotranspiration Levels Composite by Ramon Suarez, Venezuela