eo-WIMEX is a European Space Agency funded initiative proposing an innovative framework enabling you to develop, validate and use your forward and inverse electromagnetic wave interaction models through a systematic approach.
The Earth Observation (EO) wave interaction modelling domain has witnessed a proliferation of diverse forward and inverse models developed by the scientific community over the last decades.
This wealth comes with its challenges.
Specifically, a lack of systematic implementation, testing and validation confined to small datasets, and a limited integration with Artificial Intelligence techniques.
That’s where eo-WIMEX (Wave Interaction Models EXploitation) steps into the scene.
Leveraging the unprecedented amount of EO data today available and resorting where needed to GPU and parallel computing, the framework is designed to enhance the overall efficiency of your forward and inverse models ensuring a flexible environment which adapts to your needs.
eo-WIMEX is designed to host, manage and exploit any forward and inverse model.
Our Model Development Environment facilitates the development and testing of your model and guides you through a standardized containerization process of your application, for a smooth integration.
The containerization makes the framework treat your model like a “black box”, a key feature which ensures eo-WIMEX remains independent of the model type.
After the integration, your model finds its place in a searchable Model Catalogue, ready to be shared with the scientific community or just with your team.
Alongside this, you can benefit from a Versioning Control Service, empowering you to manage multiple versions of your model and its associated documentation.
eo-WIMEX connects with different local and cloud data sources and databases, including EO data, in situ data, and ad-hoc datasets. This connection helps you with the design and validation of your forward models, creating and storing large volumes of lookup tables and datacubes.
These outputs from the forward models help test and calibrate your inverse models, supporting neural networks training or simulating sensitivity of the variables over a large set of input parameters.
eo-WIMEX is designed to handle the interaction with both local and external cloud resources, allowing constant monitoring of their status as well as the scaling up and down based on the workload.
No matter where your model runs, thanks to the configuration of the computational resources as Kubernetes clusters, eo-WIMEX allows for a processing solution agnostic to the underlying hosting infrastructure.
Fill in your contact details to be invited to participate in the beta testing of the eo-WIMEX prototype.
You will help us refine the framework prototype on remote sensing over land, initially focusing on forward and inverse models for soil moisture and snow water equivalent.
