Skip to content

multi-temporal-crop-classification

Back to geoai datasets

Field Value
Description Dataset Summary
This dataset contains temporal Harmonized Landsat-Sentinel (HLS) imagery of diverse land cover and crop type classes across the contiguous United States for the year 2022. The target labels are derived from USDA's Crop Data Layer (CDL). Its primary purpose is to train and evaluate segmentation models for multi-class crop type prediction.

Dataset Structure
Each GeoTIFF covers a 224x224 pixel area at 30m resolution and contains 18 bands (6 spectral bands repeated for three time steps). Corresponding mask GeoTIFFs contain one band with pixel-level crop class labels.

Band Order
Blue (B02), Green (B03), Red (B04), NIR (B8A), SWIR1 (B11), SWIR2 (B12), repeated across three temporal observations.

Mask Classes
0: No Data, 1: Natural Vegetation, 2: Forest, 3: Corn, 4: Soybeans, 5: Wetlands, 6: Developed/Barren, 7: Open Water, 8: Winter Wheat, 9: Alfalfa, 10: Fallow/Idle Cropland, 11: Cotton, 12: Sorghum, 13: Other.

Data Splits
3,854 chips randomly split into 80% training and 20% validation sets.

Dataset Creation
Samples were drawn from USDA CDL to ensure representative coverage across CONUS. For each chip, low-cloud HLS S30 scenes between March and September 2022 were selected, reprojected to EPSG:5070, and stacked into 18-band inputs. Chips containing clouds, shadows, or missing data were filtered using Fmask quality control.
Folder /datasets/geoai/ibm-nasa-geospatial/multi-temporal-crop-classification
Discipline GeoAI / Remote Sensing / Agriculture
DOI 10.57967/hf/0955
Link Access Data
Public True
Publication Date 2025-02-11
Downloaded 2025-08-19
Data Type GeoTIFF
Dataset Size 37G
Number of Files 15511
Usage 
$ module avail
$ module load datasets
$ module load geoai/ibm-nasa-geospatial/multi-temporal-crop-classification/2025-02-11
Usage Policy Link https://choosealicense.com/licenses/cc-by-4.0/
Usage Policy This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0)
license.
The CC BY 4.0 license allows sharing, adaptation, and reuse of the dataset for both academic and commercial purposes, provided that appropriate credit is given to the original authors. Users should include attribution to the dataset creators and the IBM–NASA Geospatial team when using this dataset in publications, derived models, or applications.
Citation Cecil, M., Kordi, F., Li, H. (S.), Khallaghi, S., & Alemohammad, H. (2023). HLS Multi-Temporal Crop Classification Dataset (v1.0) [Dataset]. Hugging Face. https://huggingface.co/datasets/ibm-nasa-geospatial/multi-temporal-crop-classification https://doi.org/10.57967/hf/0955
BibTeX
📜 View BibTeX citation
@misc{ibm_nasa_geospatial_2023,
   author       = { IBM NASA Geospatial },
 title        = { multi-temporal-crop-classification (Revision 9b51700) },
   year         = 2023,
    url          = { https://huggingface.co/datasets/ibm-nasa-geospatial/multi-temporal-crop-classification },
  doi          = { 10.57967/hf/0955 },
    publisher    = { Hugging Face }
}