Earth Observation data and Artificial Intelligence in support of Journalism

31 March 2020

Anastasios Drosou, Ioannis Manakos
Centre for Research and Technology, Greece

Earth Observation data is valuable for journalist’s reports to the public. An example are the maps released in little time during or after the tsunami in Indian Ocean in 2004 or the Fukushima disaster in 2011, accompanying the verbal or text reports of theirs. Taking advantage of the improved temporal frequency and spatial cover of the Sentinel satellite sensors SnapEarth aims to assimilate latest spaceborne retrieved information to support journalists in their work in near real time.

In this context, a dedicated services’ module aims to leverage on Copernicus monitoring services, like the EMS’s (Emergency Management Service) EFAS (European Flood Awareness System) and EFFIS (European Forest Fire Information System). It will add in tandem to them the ability to exploit latest AI (Artificial Intelligence) techniques to automatically and unsupervised query through big data piles to deliver in minimum time required products. This way Petabytes of information, offered by Copernicus, is placed to the journalists reach. Journalists’ capabilities will be thus enhanced in regards to their reporting on natural or human-made disasters affecting various sizes of areas and populations.

To date, Copernicus provides information on major events via its EMS. Information provided ranges from the provision of rapid and recovery maps, for emergency response or prevention actions respectively, to early warnings for floods, droughts and forest fires. Rapid maps may be delivered on-demand via the Copernicus EMS to illustrate the pre- and post-event situation for a specific area of interest. Pre-event maps (reference maps) are based on images captured as close as possible prior to the event. Post-event maps include the first estimate product, which is a first fast assessment of the situation, and the delineation and grading products, which provide information on the extent and damage grade. Information by the EFAS providing flood forecasts and the EFFIS for fire danger forecasts may be also retrieved and used as a trigger for the SnapEarth AI queries, where useful. Within SnapEarth innovators will additionally exploit latest results developed within other H2020 research and innovation projects (e.g. ECOPOTENTIAL), which are based on forms of AI. These are techniques for flooded areas’ detection via machine learning application using optical and radar Sentinel images, and targeted services for abrupt phenological changes’ detection via phenological cycles’ assessment using Landsat and Sentinel multispectral time series images. Latter is applied for extreme events detection, such as forest fires’ or drought’s scars detection on the landscape. Workflows will be properly adjusted to the journalists’ needs and provide for a series of services via EarthPress.

Figure 1: Maps of water extent changes within Doñana National Park, with focus on the marshland area. The maps were generated from each available cloud free observation of Sentinel-2 data, where white color indicates presence of open water surface and black color indicates dry land/ vegetation. (Credit: ECOPOTENTIAL Project No 641762 - Deliverable No 4.3: EO Change Detection Modules)

The EarthPress pilot service of the SnapEarth Project aims to help the Media & Press Industry. The pilot will employ a variety of Artificial Intelligence tools, collecting data from social media, images and data from the Copernicus services and other sources, concerning earth info, in order to generate multimodal text. The text will be in natural language form, providing ready-to-publish material that the end-user can directly use.

Artificial Intelligence is a rapidly evolving field that has entered our lives in recent years giving impetus to scientific disciplines and influencing our daily lives with the technological advances it offers. Scientific fields currently utilizing AI include Healthcare and Medical Imaging Analysis, Agriculture and Farming, Self-driving Cars and many more.

Artificial Intelligence, although used in several areas, has not been exploited within the scope of earth observation tools or from the Press Industry. EarthPress aspires to fill this void. The purpose of the program is to make EO data more usable and user friendly so that the user receives the information s/he needs through a simple query, with the help of AI tools.


Leave a comment