Search papers, labs, and topics across Lattice.
The authors introduce YieldSAT, a large-scale, high-resolution multimodal dataset for crop yield prediction spanning multiple countries and crop types. The dataset contains over 12.2 million yield samples at 10m resolution, paired with multispectral satellite imagery and auxiliary environmental data across 2,173 fields. They demonstrate the dataset's utility by benchmarking deep learning models for pixel-level yield regression and propose a domain-informed deep ensemble approach to address distribution shifts.
Training better crop yield models just got easier: YieldSAT offers 12 million high-resolution, multimodal data points across diverse climate zones and crops.
Crop yield prediction requires substantial data to train scalable models. However, creating yield prediction datasets is constrained by high acquisition costs, heterogeneous data quality, and data privacy regulations. Consequently, existing datasets are scarce, low in quality, or limited to regional levels or single crop types, hindering the development of scalable data-driven solutions. In this work, we release YieldSAT, a large, high-quality, and multimodal dataset for high-resolution crop yield prediction. YieldSAT spans various climate zones across multiple countries, including Argentina, Brazil, Uruguay, and Germany, and includes major crop types, including corn, rapeseed, soybeans, and wheat, across 2,173 expert-curated fields. In total, over 12.2 million yield samples are available, each with a spatial resolution of 10 m. Each field is paired with multispectral satellite imagery, resulting in 113,555 labeled satellite images, complemented by auxiliary environmental data. We demonstrate the potential of large-scale and high-resolution crop yield prediction as a pixel regression task by comparing various deep learning models and data fusion architectures. Furthermore, we highlight open challenges arising from severe distribution shifts in the ground truth data under real-world conditions. To mitigate this, we explore a domain-informed Deep Ensemble approach that exhibits significant performance gains. The dataset is available at https://yieldsat.github.io/.
