About the Analysis Ready Data (ARD) category

One of the key objectives of providing analysis ready data (ARD) is making earth observation data ready for (e.g., time-series) analysis and making it more accessible to non EO experts. A major problem with ARD is however also to find the level of pre-processing that enables a wide range of applications, without excluding others. The harmonisation and normalisation of data is also one of the prerequisites to enable time-series analysis and for creating data cubes and making them available as interoperable collections. One of the first goals of openEO Platform is to provide two main categories of ARD, multi-spectral imagery from the optical domain and synthetic aperture radar (SAR) backscatter. Further, three different processing pipelines are available for accessing and generating ARD. For fast and convenient access, pre-computed collections of ARD that follow the CARD4L are available. For both optical and SAR data, on-demand (asynchronous batch jobs) and on-the-fly (synchronous jobs) are implemented. On-demand is thought to be used for large area pre-processing for analysis with special requirements. On-the-fly processing is used for quick access and rapid prototyping on areas with less strict requirements or smaller extent. The ARD processing capabilities in openEO Platform will also allow for user defined parameterisation, while CARD4L provides the baseline specification. This feature will allow users to generate ARD to their specific requirements e.g., including topographic normalisation of surface reflectance.

  • Please discuss the ARD Level of preprocessing.
  • What is your definition of on-the-fly processing?
  • What is your understanding of the goal in ARD processing?
  • What are the future delvelopments in ARD data preparation?