EarthSelf

Purpose

The EarthSelf-Service enables any business or user to take advantage of highly scalable cloud environments for analysis, visualization and production of Earth observation value-added services. These environments allow all users to create new value-added geographic services by exploiting the results of the EarthSignature database and combining them with their own data and other Earth observation data made available in a directly usable form (ARD: Analysis Ready Data). EarthSelf-Service offers a set of tools to facilitate the development, testing, packaging, deployment and the access to the data. A ready-to-use infrastructure, hosted on European Cloud resources, ensures the ramp-up of services and large-scale productions.

Foundations

Our solution is built on the well-known Kubernetes environment and managed by SafeScale the open source portable and secured multi cloud management solution developed by CS Group.

Kubernetes is the reference platform for the deployment of flexible, scalable and highly available services. Kubernetes environments deployed with SafeScale are optimized for security and operability to guarantee an irreproachable quality of service regardless of the cloud provider chosen.

The interactive environment is powered by Jupyter, put on steroid thanks to powerful extension such as data accessor (thanks to the CS Group EODAG open source library).

 

Service description

The proposed SnapEarth EarthSelf-Service allows on-demand provisioning of Earth observation data exploitation environment. Users will be able to leverage the SnapEarth Earth Signature database, the ease of access to satellite data and the highly scalable and robust infrastructure offered by EarthSelf Service to deploy their projects in the Cloud as secure services. Thanks to the powerful workflow engine integrated in the solution, complex processing chains can be published combining classical processing and artificial intelligence.

SafeScale

The optional SnapEarth Cloud Desktop service will provide users with a graphical environment allowing them to prototype and fine-tune various parameters of their projects. It provides all tools involved in the analysis of Earth observation data and the development of added value services. It is based on web technologies for a fluid, ergonomic and intuitive user experience without any deployment on the user infrastructure.

SnapEarth EarthSelf-Service Platforms are designed to combine seamlessly the best open source tools:

  • access to Earth observation data with eodag;
  • data mining with Jupiter, numpy/scipy, xarray, pandas...;
  • parallel and distributed computing with dask;
  • machine learning with scikit-learn, pytorch, TensorFlow;
  • Earth observation product processing with SNAP, OTB, … ;
  • scientific data visualization with PyViz tools.