Search papers, labs, and topics across Lattice.
This chapter explores the transition from traditional Earth observation models to Earth embeddings, which are compact vectors summarizing satellite imagery features for easier analysis. By comparing various types of Earth embeddings, including implicit location encoders and explicit patch products, the authors highlight their applications in land cover mapping, ecological modeling, and socioeconomic prediction. Key findings indicate that while embeddings can enhance feature analysis, their effectiveness can be limited by pooling, fusion, or spatial transfer methods, with practical case studies illustrating their utility in similarity search and land cover mapping.
Earth embeddings could revolutionize satellite data analysis by enabling users to leverage compact feature vectors without the overhead of processing raw imagery.
Earth observation is moving from foundation models that users must run themselves toward embedding products that package model feature outputs as reusable data without needing to download and process the imagery used to generate them. Earth embeddings are vectors that summarize locations, image patches, or pixels, letting users analyze compact features instead of repeatedly training or running large models on raw satellite imagery. This chapter explains the main types of Earth embeddings, from implicit location encoders to explicit patch and pixel products, and compares their coverage, resolution, dimensionality, storage cost, licenses, and reproducibility. We review their use in land cover and crop mapping, ecological and hazard modeling, socioeconomic prediction, and semantic search, with evidence on when embeddings improve on conventional features and when pooling, fusion, or spatial transfer limit performance. Two case studies show practical workflows for similarity search and land cover mapping. We close with guidance for choosing, evaluating, storing, compressing, and publishing embeddings, and with open problems in oceanic and atmospheric coverage, uncertainty, and benchmarking.
