A processing system to generate custom cropland and crop type maps
Ten months after the start of WorldCereal Phase II, the project team launched the first version of the new WorldCereal cloud-based processing system for cropland and crop type mapping at field scale.
During a second user webinar the WorldCereal team explained more about the different components of the processing system and showed how you can make use of the system to generate custom cropland and crop type maps at field scale. This first version of the processing module allows users to:
- Apply a pre-trained cropland detection model on a region and year of interest. This model has been trained on a global land cover dataset, spanning multiple years and should be readily transferable both in space and time. We are of course continuously updating this model as more in-situ reference data is being added to the system, through the Reference Data Module.
-
Train and apply your own cropland detection model based on your definition, for your region and year of interest. For this, you can make use of all public land cover training data currently present in the system.
-
Train and apply your own crop type detection model for your region, season and crop types of interest. Training of models can be done using the publicly available crop type data currently ingested in the system. Soon however, we will release the next version, allowing you to effectively use your own (private or public) in-situ reference data as well.
Access the Jupyter Notebooks
Above functionalities are accessible through a series of user-friendly Jupyter Notebooks, hosted on our WorldCereal GitHub repository. Consult the README file of the repository to learn how to access and run these notebooks without the need to install anything!
Additional system components and resources
Aside from the Processing Module, the WorldCereal system also offers additional system components and resources. View the webinar recording to learn more, explore the data and resources, and join the WorldCereal community!
- WorldCereal documentation portal for more background and information about the project and the processing system.
- Reference Data Module allowing you to explore and download publicly available in-situ reference data, as well as upload and harmonize your own (private) data through an intuitive user interface.
- Visualization and Dissemination Module allowing you to explore the global WorldCereal products generated for 2021.
Do you have any other questions? Please engage with us directly through our brand new user forum (public) or contact us.