TL;DR
Get audio and creator gear delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
OlmoEarth Studio has launched a feature allowing users to generate and export custom embedding vectors from satellite data. This enhances tasks like similarity search and land cover classification, though performance and access details are still emerging.
OlmoEarth Studio has introduced a new capability that allows users to generate and export custom embedding vectors from satellite imagery based on specific regions, timeframes, and sensors. This development offers a streamlined approach for Earth observation tasks such as similarity searches and land-cover classification, without requiring users to train full models. The feature is now available through the Studio platform, with access requests currently open.
The new feature in OlmoEarth Studio supports on-demand computation of embedding vectors for selected geographic areas, dates, resolutions, and satellite sources like Sentinel-2 and Sentinel-1. Users can define their area of interest by drawing polygons or uploading data, after which the platform manages imagery acquisition and tiling. The system offers three encoder variants: Nano (128 dimensions), Tiny (192 dimensions), and Base (768 dimensions), with larger models requiring more computing resources.
The embeddings are delivered as Cloud-Optimized GeoTIFFs, with each band representing an embedding dimension. Values are stored as signed 8-bit integers, with a published dequantization function available to recover floating-point vectors. Since the computation is performed on demand, the output reflects the specific geographic and temporal parameters chosen by the user. The platform supports multiple use cases, including similarity search, clustering, and few-shot land-cover classification, although the performance across different environments remains to be fully validated.
Implications for Earth Observation and AI Applications
This development simplifies access to satellite data representations, enabling researchers and developers to perform advanced analyses such as similarity searches and land-cover segmentation more efficiently. By providing pre-computed, customizable embeddings, OlmoEarth reduces the need for extensive model training, lowering barriers for smaller organizations and individual researchers. However, the platform’s performance across diverse climates and sensors has not yet been comprehensively validated, and access terms are still being defined.
Products Worth Considering
Samsung Galaxy Tab A7 Lite 8.7 32GB Wi‑Fi Android Tablet – Long‑Lasting Battery, Compact Slim Design, Sturdy Metal Frame (US Version)
Apple MacBook Air 15-inch with M5 chip, 24GB RAM, 512GB SSD, Midnight
Samsung Galaxy Tab A11 (2026) – 8.7 90Hz Android Tablet, 64GB/4GB Wi‑Fi, Octa-core MediaTek Helio G99, Dual Dolby Speakers, 20W Fast Charger
As an affiliate, we earn on qualifying purchases.
Background on OlmoEarth and Satellite Embeddings
OlmoEarth is an open-source project offering foundation models for Earth observation data, with publicly available code, weights, and research papers. Its platform enables users to generate satellite image embeddings tailored to specific regions and time periods, supporting applications like similarity search, clustering, and segmentation. Prior to this update, users relied on full model training or external processing to obtain such representations, which could be resource-intensive and less flexible.
The recent addition of on-demand embedding exports marks a significant step toward making advanced satellite data analysis more accessible and customizable, aligning with broader trends in AI-driven Earth observation.
“OlmoEarth Studio now lets you compute and export embedding vectors tailored to specific regions and timeframes.”
— Thorsten Meyer, OlmoEarth team
Unresolved Questions About Performance and Access
It is not yet clear how well the embeddings perform across different climates, sensors, and real-world tasks. The announcement does not specify processing times, pricing, geographic restrictions, or detailed validation results. The effectiveness of the embeddings for operational or large-scale applications remains to be independently verified.
Next Steps for Users and Developers
Interested users can request access to the Studio platform, with availability likely expanding as the team clarifies access terms. Further validation studies and performance benchmarks are expected to be published, providing clearer guidance on use cases. Developers and researchers are encouraged to experiment with the open-source models to assess suitability for their specific needs.
Key Questions
What is new about OlmoEarth Studio’s latest update?
It now supports on-demand generation and export of satellite image embedding vectors for specific regions, timeframes, and sensors, streamlining Earth observation analysis.
What formats are the embeddings exported in?
They are delivered as Cloud-Optimized GeoTIFF files with one band per embedding dimension, stored as signed 8-bit integers. Floating-point vectors can be recovered using a published dequantization function.
What applications can these embeddings support?
Potential uses include similarity search, clustering, land-cover classification, and unsupervised exploration, though performance varies depending on data and task specifics.
Is the OlmoEarth model publicly available?
Yes, the source code, model weights, and research paper are publicly accessible, allowing independent computation of embeddings outside the Studio platform.
What are the limitations or uncertainties of this new feature?
Details about processing times, costs, and performance validation are still unclear. Its effectiveness across different environments and operational scenarios remains to be confirmed.
Source: ThorstenMeyerAI.com
