Abstract
Anyone who wants to use high-resolution satellite imagery such as from the Landsat or Copernicus programs over larger areas or over larger time spans knows that downloading the required data is in most cases no longer really feasible. Analysing the data in the cloud requires the analyst to choose between two options: (i) renting on or more virtual machines in a cloud that provides the data, install the software, find all the required tiles and carry out the analysis, or (ii) sign up for Google Earth Engine (GEE, ™) and carry out the analysis there. The first option provides the ultimate flexibility but requires a lot of technical skills. The
second is easy, but also has a number of limitations, ranging from the application of user-defined algorithms, the requirement of using open source software, or the constraints imposed by GEE’s end-user-license-agreement.
The openEO Platform provides a middle ground between the two extremes: like GEE it provides a high-level API where problems can be expressed as operations on images collections, but unlike GEE it provides access to any compatible cloud provider, many of which build 100% on open source software and many of which allow the user to execute user-defined functions, custom code (in Python or R) that specifies the computations that the user wants to carry out e.g. on pixel time series or over a series of spectral bands. This training workshop will introduce the concepts of openEO (the API, the processes it specifies) as well as the tooling that is available around it (the clients in R, Python, QGIS, or JavaScript; the back-end connectors to cloud platforms; the hub and the validator). It will also provide a number of examples with which participants can experiment, using a free tier access of openEO platform (operated by EODC, VITO and Sinergise).