Logo Lanfrica
  • Accueil
  • Atlas
  • Analyses
  • Documentation
  • Sign in

© 2026 Lanfrica. Tous droits réservés. Tous les droits d'auteur des ressources affichées sur le site Web Lanfrica appartiennent aux détenteurs de droits d'auteur d'origine, sauf indication contraire explicite.

WorldCereal/worldcereal-classification: WorldCereal classification module v2.6.0

Domaine:

agriculturegeospatial

Type de record:

software
Créateur:
KriJerVinHan
Éditeur:
Zenodo
Hôte:avatar

This release marks the transition from a classification system built on Presto embeddings + CatBoost classifier towards a seasonal torch head directly trained on the embeddings.

What's Changed

  • Import default seed from prometheo by @kvantricht in github.com
  • fix private ref data extractions by @jdegerickx in github.com
  • Add the option to have controlled val and test splits by @kvantricht in github.com
  • Solve STAC merge conflicts by @cbutsko in github.com
  • meteo has to be spatially explicit for prometheo update by @kvantricht in github.com
  • Add embeddings as a worldcereal product by @kvantricht in github.com
  • Minor improvements by @jdegerickx in github.com
  • World cereal description minor change by @tho-hole in github.com
  • Expose and document time_explicit in notebook functionality by @kvantricht in github.com
  • Add an example notebook on embeddings generation by @kvantricht in github.com
  • Configurable job options args for patch-to-point by @cbutsko in github.com
  • Added inputs collection based on job manager by @kvantricht in github.com
  • Notebooks by @jdegerickx in github.com
  • process_parquet bugs by @cbutsko in github.com
  • Correct import of MIN_EDGE_BUFFER by @cbutsko in github.com
  • Add human labels to training_df by @cbutsko in github.com
  • String label columns to training_df by @cbutsko in github.com
  • Pin patch-to-point workflow to python 3.8 until it works on python 3.11 by @VincentVerelst in github.com
  • Empty DF handling by process_parquet by @cbutsko in github.com
  • Notebook updates by @jdegerickx in github.com
  • Local inference + extraction of preprocessed inputs by @jdegerickx in github.com
  • add ref to preprocessed inputs notebook by @jdegerickx in github.com
  • Fix dependencies by @jdegerickx in github.com
  • Realistic extractions dates by @cbutsko in github.com
  • Exposing frequency argument in collect_inputs by @cbutsko in github.com
  • add explicit scikit-image dependency in udf_distance_to_cloud.py #462 by @VincentVerelst in github.com
  • Updated public datasets alignment by @cbutsko in github.com
  • Add trimming functionality process_parquet by @cbutsko in github.com
  • Stabilize job options for patch extractions by @cbutsko in github.com
  • fix old versions of inference UDPs to python38 image on openeo by @VincentVerelst in github.com
  • Update data loading for finetuning by @cbutsko in github.com
  • Making on-disk-db robust to new columns by @cbutsko in github.com
  • make printing of tables nicer using tabulate by @jdegerickx in github.com
  • Sample extractions by @jdegerickx in github.com
  • Season performance by @jdegerickx in github.com
  • Allow to override presto URL by @kvantricht in github.com
  • New seasons by @jdegerickx in github.com
  • Local inference by @jdegerickx in github.com
  • Improved landcover training by @kvantricht in github.com
  • Minor notebook updates by @jdegerickx in github.com
  • limit processing period to end of sept by @jdegerickx in github.com
  • Fix private extract by @jdegerickx in github.com
  • Update local inference by @jdegerickx in github.com
  • Un handbook by @jdegerickx in github.com
  • Notebooks zambia by @jdegerickx in github.com
  • added minor documentation by @jdegerickx in github.com
  • crop type notebooks zmb by @jdegerickx in github.com
  • import notebook utils by @jdegerickx in github.com
  • Un handbook by @jdegerickx in github.com
  • Unhandbook by @jdegerickx in github.com
  • Fixes datasets and models by @kvantricht in github.com
  • Combined udf by @HansVRP in github.com
  • add test dataset download by @jdegerickx in github.com
  • Updating prometheo wheel version in inference by @cbutsko in github.com
  • Update croptype udp by @VincentVerelst in github.com
  • Fix private extractions ntb by @jdegerickx in github.com
  • explicitly reproject early with custom target_epsg by @kvantricht in github.com
  • Local inf fixes by @jdegerickx in github.com
  • Artifactory to s3 by @VincentVerelst in github.com
  • fix workflow when keep class probabilities is False by @jdegerickx in github.com
  • Patch extractions updates by @cbutsko in github.com
  • Default value of extract_value should be 0 by @kvantricht in github.com
  • UDP performance: request more cpu by @jdries in github.com
  • Change inference output dtype to uint8 by @VincentVerelst in github.com
  • set default tile size to 50 and update documentation by @jdegerickx in github.com
  • Add CRS to queried extractions by @kvantricht in github.com
  • Added forced reindexing before trimming by @cbutsko in github.com
  • urgent fix to UDP by @HansVRP in github.com
  • Validate Presto backbone checkpoint consistency for seasonal head overrides by @kvantricht in github.com
  • 324 migrate job manager by @VincentVerelst in github.com
  • Updated dependencies by @kvantricht in github.com
  • process_parquet fixes by @cbutsko in github.com
  • Seasonal head finetuning by @kvantricht in github.com
  • Extractions app by @jdegerickx in github.com
  • Various fixes by @jdegerickx in github.com
  • separate cdse connection utilities in apps by @jdegerickx in github.com
  • Updates to local inference by @cbutsko in github.com
  • Training app customization by @jdegerickx in github.com
  • Kenya updates by @jdegerickx in github.com
  • Notebook upd by @jdegerickx in github.com
  • Notebook updates by @jdegerickx in github.com
  • fix filtering ref_id in query_private by @jdegerickx in github.com
  • fix: avoid deprecated references by @pvbouwel in github.com
  • update meteo composite STAC API URLs to newest versions #539 by @VincentVerelst in github.com
  • fix pip install instructions by @jdegerickx in github.com
  • Allow custom dem by @kvantricht in github.com
  • Embeddings from seasonal model by @jdegerickx in github.com
  • Balanced train splits by @cbutsko in github.com
  • Add option to export global embeddings and auxiliary NDVI by @kvantricht in github.com
  • update to latest version of cropcalendars by @jdegerickx in github.com
  • Expose S2 masking parameters by @cbutsko in github.com
  • Patch-to-point heavy job split by @cbutsko in github.com
  • Local inference improvements by @cbutsko in github.com
  • Multiutm job dataframe creation by @cbutsko in github.com
  • New release by @jdegerickx in github.com

New Contributors

  • @tho-hole made their first contribution in github.com
  • @HansVRP made their first contribution in github.com
  • @pvbouwel made their first contribution in github.com

Full Changelog: github.com

Visit

doi.org

Licenses

info:eu-repo/semantics/openAccessMIT Licensehttps://opensource.org/licenses/MIT

Similaires

Global crop calendars of maize and wheat in the framework of the WorldCereal project

Global crop calendars of maize and wheat in the framework of the WorldCereal project

Crop calendars provide valuable information on the timing of important stages of crop development