Copernicus data are becoming tremendous in volume & quality. Copernicus services are focusing on professional services, like the Emergency Management Service focusing on providing information on major events to the rescue and security forces. In order to increase awareness of the Copernicus services to the EU citizens, there is a need to provide added value data to the Media Industry, rendering them to major demonstrators of the success of Copernicus to the citizens.
In this direction, the impact of the SnapEarth project will be enhanced through services for combinatorial EO data retrieval, data correlation between EO and news information will be developed.
In particular, the EarthPress pilot aims to tackle these issues by delivering value-added products to editors & journalists allowing them to enrich content of their publications & articles. These contents, tailored to specific user needs, will be notably based on EO data (e.g. disaster extent maps, background imagery, aggregated information from social media, blogs & websites, etc.) on a certain topic.
The EarthPress pilot will based on:
- a self-service Web portal simplifying the use of such services & decision making.
- real-time value-added data to be used within the processing chains
- support from both processing framework (i.e. SafeScale) and the computational infrastructure deployed on future Copernicus DIASs.
The EarthPress pilot will support the Media & Press Industry with a set of valuable tools & services for Εarth-related Εvent reporting and facilitate the composition of News articles. Some of the innovations/modules that the platform will introduce/utilize are listed below:
- Adaptive Profile generation & Management: Provided the fact that in the Press industry, editors & journalists focus on various topics (e.g. international, business, local, etc.), their profiles must be characterized before any information is delivered.
- Breaking News managment: This module will monitor information related to extreme events.
- It will render the main Data Production module. Being accessible through dedicated APIs, the technology of QWANT’s search engine will allow EarthPress to have a real-time access to all available information (e.g. news, pictures, social web, IoT, etc.) without advertisements.
- Earthbot: This module implements a central technology for this pilot, offering an AI-based bot that will be capable of extracting multi-media data, classify & correlate information, generate additional meta-queries and illustrate the results.
- Composition: The EarthBot will deliver an information file fitting with the journalist’s profile, facilitating the production of a well-fitting article. Typical services will include language translation, text summarization & text enhancement from multimedia data through state-of-the-art AI technologies. Approaches dealing with AI models able to process different modalities (e.g. text, image, audio etc.) in a unified matter and to perform many tasks at once, from speech recognition, and image classification and captioning, to image-to-text/text-to-text translation will also be considered for possible utilization and extension.
The idea is that the final article will be trimmed & finalized by the journalist themselves, within the given layout constraints.
A typical case study will focus on the reporting of natural disasters affecting geographic areas that can be captured by Earth observation satellite and interpret the “before-after” depiction of the disaster on an impacted area, enriched with social networks information and photographs.