AgriFieldNet India: Crop Type Labels for Smallholder Fields with Sentinel-2 Imagery

Provider
Radiant Earth Foundation and IDinsight
License
CC-BY-4.0 (source)
Format
JSON (in sampled files)
Stored at
Azure Blob Storage, https://radiantearth.blob.core.windows.net/mlhub/ref_agrifieldnet_competition_v1
Access
Free and public. No account with the data provider is needed.
DOI
10.34911/rdnt.wu92p1
Category
Earth Observation & Geospatial
Source
https://doi.org/10.34911/rdnt.wu92p1

About this dataset

Crop type labels for smallholder agricultural fields in Uttar Pradesh, Rajasthan, Odisha and Bihar in northern India, collected on the ground by IDinsight's Data on Demand team and curated by Radiant Earth Foundation for the AgriFieldNet India competition, funded through the ECAAS initiative of the Bill & Melinda Gates Foundation. There are 13 classes: fallow land and 12 crops (wheat, mustard, lentil, green pea, sugarcane, garlic, maize, gram, coriander, potato, berseem and rice), with fields that are often well under one hectare. Labels are rasterised to 256 x 256 pixel chips aligned with Sentinel-2 L2A imagery at 10 m. Measured 2026-11-15: 17,644 GeoTIFFs (0.33 GB). source/ holds 12-band Sentinel-2 chips (B01 to B12) per chip id, train_labels/ holds field id and crop class rasters for the training chips and test_labels/ the field id rasters for the test chips, with STAC JSON alongside. Load with rasterio or the radiant-mlhub and torchgeo AgriFieldNet loaders. Released under CC BY 4.0; cite Radiant Earth Foundation and IDinsight (2022), AgriFieldNet Competition Dataset, doi:10.34911/rdnt.wu92p1.

How to access this dataset in Python

import adlfs
fs = adlfs.AzureBlobFileSystem(account_name='radiantearth', anon=True)
fs.ls('mlhub/ref_agrifieldnet_competition_v1/')[:10]

Lists the files with anonymous access. Runs on any machine, with no cloud account.

Data structure

File types in sampled files: JSON

Files

Example file: .source/metadata.json

Analyze this dataset without downloading it

On Scigantic, AgriFieldNet India: Crop Type Labels for Smallholder Fields with Sentinel-2 Imagery opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Earth Observation & Geospatial datasets