COLD-CI: A large-scale very high-resolution label polygon dataset for cocoa and non-cocoa classification in Côte d'Ivoire

COLD-CI consists of 123,736 vector polygons corresponding to a total labelled area of 5,996 km², including 58,107 cocoa polygons (1,788 km²) and 65,629 background polygons (4,208 km²). Polygon label candidates were generated through automated filtering and integration of the West Africa Cocoa polygon dataset (WAC; Schneider et al., 2023) and multiple external thematic datasets. These candidates were subsequently refined, validated, and complemented through systematic visual interpretation, manual correction, and digitisation using very high-resolution (0.5 m) satellite imagery. The resulting label polygons capture cocoa planted areas and associated fine-scale internal heterogeneity, as well as a wide range of non-cocoa land-cover types. This Zenodo release provides open access to the validated background (non-cocoa) polygons and the footprints of the Pléiades image scenes used during dataset construction. The full COLD-CI dataset, including cocoa polygons, is available separately through the World Resources Institute (WRI) under a Data Sharing Agreement. The dataset is intended to support Earth Observation model development, validation, and benchmarking across a wide range of spatial resolutions.