TELESPAZIO VEGA UK LIMITED (GB)
The EASEQC project aimed at expanding the use of AI/ML for quality control of EO products. The traditional approach to quality control, usually involving deterministic models together with considerable manual intervention, is no longer feasible given increasing data volumes of EO data archives. ML/AI has potential to make the process of quality control more efficient. EASEQC focused on the development of semi-supervised ML models for detection anomalies in EO products. This entailed that models can be trained with limited training data and that a model is capable of identifying generally anomalous data products i.e. different anomalies can be detected by the same model. The service has been implemented in a cloud environment and is accessible via an API.
Overall, the outcome of the project has seen significant steps made towards the establishment of an operational Ease QC service. Further work is still required to improve the ML models, but the infrastructure successfully developed by the project both with respect to the development of the ML models, and their deployment / operation alongside the data (be that on the cloud or otherwise) is an extremely significant development with respect to the long term objectives of the Ease QC team.