Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a weakly supervised approach for identifying and ranking potential dairy farm sites using seasonal satellite imagery and open map data. By leveraging a Barlow Twins encoder to learn embeddings from multi-season image tiles and integrating weak OpenStreetMap priors, the method effectively ranks candidate clusters based on proximity to known farm features and seasonal vegetation patterns. The results demonstrate that the approach can yield high precision in identifying candidate sites, achieving 0.60 precision within 500 meters of held-out farm features, thus streamlining the farm discovery process from extensive satellite datasets.
Seasonal representation learning combined with weak geographic priors can pinpoint dairy farm sites with surprising precision, even from incomplete labels.
Farm site discovery from satellite imagery is a spatiotemporal candidate ranking problem because farm evidence is distributed across pasture, field boundaries, roads, buildings, and seasonal vegetation patterns. Direct farm labels are often incomplete, which makes fully supervised detection difficult. This paper proposes a weakly supervised pipeline for ranking dairy farm candidate clusters from seasonal Sentinel imagery and open map priors. The method uses aligned spring, summer, and autumn image tiles from County Cork, Ireland, with spectral bands, vegetation indices, built area indices, and a pasture channel. A Barlow Twins encoder learns multi-season tile embeddings without farm labels. In parallel, weak OpenStreetMap farm priors are split into a prior and a held-out set. Prior features support a rule-based tile score that combines farm proximity, seasonal pasture evidence, and summer greenness, while held-out features are reserved only for proxy evaluation. The rule score is smoothed over a spatial representation graph using geographic proximity and embedding similarity, and high-scoring tiles are grouped into ranked candidate clusters. From 26,722 valid tiles, the main run selects 535 high-confidence tiles and forms 71 candidate clusters. The top 5 clusters achieve 0.60 precision within 500 m and 0.80 precision within 1000 m of held-out OpenStreetMap farm features. The top 10 clusters achieve 0.40 precision within 500 m and 0.80 precision within 1000 m. The results show that seasonal representation learning and weak geographic priors can reduce large satellite image collections into compact candidate sets for human review.