CAFOSat: A Strongly Annotated Dataset for Infrastructure-Aware CAFO Mapping Using High-Resolution Imagery
Researchers have developed CAFOSat, a new dataset designed to improve the mapping of Concentrated Animal Feeding Operations (CAFOs) using high-resolution imagery. This dataset integrates satellite imagery with refined annotations of infrastructure like barns and manure ponds, addressing challenges posed by inconsistent data and layouts. CAFOSat includes over 45,000 image patches across 20 states and serves as a benchmark for evaluating various AI models, including convolutional, transformer-based, and vision-language architectures. AI
IMPACT Provides a benchmark for AI models in agricultural monitoring and disease surveillance.