Today sees the launch of a new, free and open, service, the CEOS-Product Validation Platform (CEOS.org/PVP), to help quality assure and harmonise the world’s optical Earth Observation satellite data and derived information. This service will facilitate increased utility and integration of the world’s space assets – commercial and public, enabling more reliable decision making and opportunities for economic growth through innovative services, from this dynamic sector. The beta version of this service has been pre-populated with data from a few sensors with the expectation that this will increase over the coming months and as the service itself expands in functionality.
Robust independent assessment and consequently correction, of any bias, is a necessary first step to allow satellite operators, public and commercial, and their customers to readily integrate data/information derived from different satellites into knowledge driven actions and services, particularly those where time dependent change is important. Whilst every effort is made to ensure data from satellites is accurate, the shock of launch and harshness of space generally leads to small but significant changes in performance which need to be assessed and corrected before the data can be exploited.
The CEOS-Product Validation Platform (CEOS-PVP) has been created by scientists at NPL on behalf of (CEOS | Committee on Earth Observation Satellites) the coordinating body of the world’s space agencies for Earth data, currently chaired by the UK. CEOS supports both governmental members but also the commercial ‘smallsat/newspace’ community to improve and evidence the quality of their data. Small satellites, generally, do not have the capacity to host on board calibration systems and rely solely on calibration/validation against externally viewed ‘references’ such as deserts, whilst the larger satellites can make use of both.
Over the last few decades, a set of ‘recommended references’ have been created: specific deserts, instrumented sites e.g. RadCalNet and the moon to serve as common infrastructure for the community. These are described and accessible through the CEOS Cal/Val portal.
Using the recently created UK EODH (UK Earth Observation Data Hub) as a host, NPL has built the CEOS-PVP. The PVP consists of a repository where satellite operators can provide images from their satellites, regularly collected over a subset of CEOS references, radiometric and geometric, as an accessible archive of evidence. In addition, for radiometry (reflectance/radiance) a RadVAL tool has been created to enable easy visualisation of a sensors’ performance compared to a specified reference as a function of time. To facilitate interoperability and harmonisation between sensors, NPL has created a ‘virtual reference’ from a combination of comparisons over different references. This virtual reference will serve as an independent baseline against which any sensor can report its performance.
The CEOS-PVP and in particular the RadVal tool, will enable the global EO community, data providers and users, to readily evidence and visualise the self-consistency and relative bias (compared to individual reference sites and /or the virtual reference) of contributing satellite sensors. This will help make interoperable Analysis Ready Data (ARD) from different sensors to be combined and substituted into services and digital twins. Looking to the future, it will allow full utilisation of Artificial Intelligence techniques to exploit the full wealth of archived EO data to meet societies increasing demands. The UK EODH, for example, will support and enable use of these trusted CEOS-PVP tools within its unique quality assurance services. In the longer term, with the launch of satellite reference missions, SITSat (SI-Traceable Satellites) such as the ESA - TRUTHS mission, led by the UK, a robust anchor value, ‘truth’, can be assigned to the virtual reference. This will further increase consistency between sensors and also result in reduced uncertainty, improved confidence and trust in the data delivered by those sensors. The latter of particular importance where small signals and/or long-term trends need to be unequivocally detected e.g. to assess progress resulting from climate action or where financial risk/litigation may become an issue.
13 Aug 2025