# On the Generalizability of Foundation Models for Crop Type Mapping

arXiv: https://arxiv.org/abs/2409.09451 · HTML: https://mlsanity.com/paper/2409.09451/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 49 references that are in the map.
- First posted on arXiv 14 Sep 2024.

## Builds on (most cited first, 17 of 17)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Improved Baselines with Momentum Contrastive Learning](https://mlsanity.com/paper/2003.04297/) (2020)
- [Seasonal Contrast: Unsupervised Pre-Training from Uncurated Remote Sensing Data](https://mlsanity.com/paper/2103.16607/) (2021)
- [SatlasPretrain: A Large-Scale Dataset for Remote Sensing Image Understanding](https://mlsanity.com/paper/2211.15660/) (2022)
- [Lightweight, Pre-trained Transformers for Remote Sensing Timeseries](https://mlsanity.com/paper/2304.14065/) (2023)
- [TorchGeo: Deep Learning With Geospatial Data](https://mlsanity.com/paper/2111.08872/) (2021)
- [Remote Sensing Vision-Language Foundation Models without Annotations via Ground Remote Alignment](https://mlsanity.com/paper/2312.06960/) (2023)
- [Multi-Modal Temporal Attention Models for Crop Mapping from Satellite Time Series](https://mlsanity.com/paper/2112.07558/) (2021)
- [SSL4EO-L: Datasets and Foundation Models for Landsat Imagery](https://mlsanity.com/paper/2306.09424/) (2023)
- [Crop Yield Prediction Using Deep Neural Networks](https://mlsanity.com/paper/1902.02860/) (2019)
- [A GNN-RNN Approach for Harnessing Geospatial and Temporal Information: Application to Crop Yield Prediction](https://mlsanity.com/paper/2111.08900/) (2021)
- [Fewshot learning on global multimodal embeddings for earth observation tasks](https://mlsanity.com/paper/2310.00119/) (2023)
- [Towards Global Crop Maps with Transfer Learning](https://mlsanity.com/paper/2211.04755/) (2022)

## Built on by (most cited first, 5 of 5)

- [PANGAEA: A Global and Inclusive Benchmark for Geospatial Foundation Models](https://mlsanity.com/paper/2412.04204/) (2024)
- [From General to Specialized: The Need for Foundational Models in Agriculture](https://mlsanity.com/paper/2507.05390/) (2025)
- [Farm-Level, In-Season Crop Identification for India](https://mlsanity.com/paper/2507.02972/) (2025)
- [Recoverable Geographic Location Information in Earth-Observation Embeddings](https://mlsanity.com/paper/2609.29151/) (2026)
- [Transferability and operational reliability of a Prithvi crop classification foundation model under phenological and geographic shift across three continents](https://mlsanity.com/paper/2610.08810/) (2026)

## Abstract (arXiv)

Foundation models pre-trained using self-supervised learning have shown powerful transfer learning capabilities on various downstream tasks, including language understanding, text generation, and image recognition. The Earth observation (EO) field has produced several foundation models pre-trained directly on multispectral satellite imagery for applications like precision agriculture, wildfire and drought monitoring, and natural disaster response. However, few studies have investigated the ability of these models to generalize to new geographic locations, and potential concerns of geospatial bias -- models trained on data-rich developed nations not transferring well to data-scarce developing nations -- remain. We evaluate three popular EO foundation models, SSL4EO-S12, SatlasPretrain, and ImageNet, on five crop classification datasets across five continents. Results show that pre-trained weights designed explicitly for Sentinel-2, such as SSL4EO-S12, outperform general pre-trained weights like ImageNet. While only 100 labeled images are sufficient for achieving high overall accuracy, 900 images are required to mitigate class imbalance and improve average accuracy.