In general, infrastructures such as cyberinfrastructures or Spatial Data Infrastructures (SDIs), allow information sharing across distributed infrastructures and communities. SDIs have gradually changed from a pool of authoritative data shared using standardized web services to a pool where the authoritative data co-exist with data collected by volunteers and different sensors. Many efforts were dedicated to data documentation, to improving the catalogues searching techniques by means of, for example, thesauri and to sharing these data using standardized web services such as Web Map Service, Web Feature Service or Web Coverage Service. Cloud computing technologies played an important role in the implementation of sustainable SDIs due to their ability to provide on-demand computational and storage capacities over the Internet. In this way, users can easily search, find and use data shared across different online platforms.
More specifically, infrastructures for image processing and analysis refer to the physical and organizational facilities that allow the storage, analysis and management of the available data and products. Traditionally, this infrastructure formed a digital image processing system consisting of computer hardware with special-purpose image processing software, and peripheral input-output devices (e.g. CD or DVD drives, internet access, printers/plotters). In recent years, Earth observation is undergoing a shift to online processing making use of data cubes and vast image archives, e.g. NSF EarthCube or Digital Earth Australia, the Swiss Data Cube, the EarthServer, the E-sensing platform or the Google Earth Engine. Available infrastructures aim at sharing remote sensing data and derived products following the FAIR metrics: Findable (F), Accessible (A), Interoperable (I), Reusable (R). Thus, remote sensing data have to be documented using metadata that support FAIR data principles as follows: (1) Findable: remote sensing data are findable through data documentation, i.e. metadata, that needs to include a unique identifier of the described data. Metadata can be stored in a catalog compliant to one of the available data cataloging standards such as the SpatioTemporal Asset Catalog (STAC) compliant catalog; (2) Accessible: all data have to be openly accessible and shared using interoperable formats that allow users to find, access and reuse them; (3) Interoperable: different standards, e.g. STAC specification, have to be used to document remote sensing data; (4) Reusable: metadata have to be comprehensive enough to allow users not only to assess the fitness for purpose (e.g. lineage) but also to provide them information about how to access the generated data.