To support this challenge, we’ve curated a dataset from specific areas of interest (AOIs), ensuring that each AOI is exclusively in either the training or testing set.

• For each pixel within an AOI, you'll receive 13 bands of Sentinel-2 data, captured over a 3-year period.
• The inclusion of 3 years of data reflects the growth cycle of cocoa crops, plam plantations and forests which mature over multiple years, unlike common annual crops.

Your objective is to predict whether a given pixel corresponds to cocoa crops, palm plantation or forest cover.

For additional information about Sentinel-2 bands and their applications, refer to this guide: Sentinel-2 Bands and Combinations.