OlmoEarth Studio Adds Custom Embedding Exports for Earth Observation Analysis

OlmoEarth Studio Adds Custom Embedding Exports for Earth Observation Analysis OlmoEarth Studio now supports generating and exporting custom embedding vectors derived from open-source OlmoEarth foundation models. Delivered as Cloud-Optimized GeoTIFFs (COGs), these compact representations allow users to perform downstream geospatial tasks such as similarity search, segmentation, change detection, and unsupervised landscape analysis. Users can configure on-demand embedding exports by area of interest, time span (1–12 months), resolution (10m to 80m), imagery inputs (Sentinel-1 and/or Sentinel-2), and encoder size (Nano, Tiny, or Base). Vectors are exported as signed 8-bit integers (int8) spanning -127 to +127 across individual raster bands. Demonstrated downstream applications include zero-label similarity search, few-shot segmentation (achieving 0.84 weighted F1 with 60 labeled pixels), change detection via cosine distance, and PCA false-color exploration. Beyond frozen embeddings, OlmoEarth Studio supports supervised fine-tuning, though embedding quality remains bounded by underlying imagery issues like cloud cover and atmospheric artifacts. On-Demand Generation and Specifications OlmoEarth Studio allows users to compute customized raster embeddings instead of relying on static pre-computed global archives. Users define an area of interest, temporal spans between 1 and 12 months, pixel resolutions from 10 to 80 meters, and imagery sources including Sentinel-2 L2A and Sentinel-1 RTC. Model architectures range from Nano (128 dimensions, 1.4M parameters) and Tiny (192 dimensions, 6.2M parameters) to Base (768 dimensions, 89M parameters). Outputs are packaged as multi-band COGs containing signed 8-bit integers ranging from -127 to +127, with -128 reserved for nodata. Downstream Geospatial Applications Frozen embeddings enable multiple downstream tasks with minimal computation. In similarity searches, cosine similarity between a query pixel or window and surrounding areas distinguishes built environments from irrigated crops without prior labels. In few-shot segmentation over Ca Mau, Vietnam, a logistic regression trained on 60 labeled pixels achieved a 0.84 weighted F1 score across mangrove, water, and other classes. Additionally, comparing monthly embeddings across years via per-pixel cosine distance accurately mapped the Park Fire burn scar in California, while Principal Component Analysis reduced features to three dimensions to map Dutch agricultural parcels without supervised guidance. Fine-Tuning and Data Limitations While frozen embeddings serve as a lightweight entry point compatible with standard geospatial libraries like GDAL, QGIS, and rasterio, OlmoEarth Studio also supports supervised fine-tuning (SFT) to train task-specific heads when higher accuracy is required. Users are cautioned that vector quality remains reliant on the source imagery; composite periods afflicted by persistent cloud cover, atmospheric artifacts, or missing observations can degrade embedding performance.

OlmoEarth Studio支持自定义导出地球观测嵌入向量 艾伦人工智能研究所（AllenAI）在OlmoEarth Studio中推出了自定义计算并导出嵌入向量（embeddings）的功能，用户可将开源地球观测基础模型生成的紧凑数值表示导出为标准云优化GeoTIFF格式，以用于多种下游分析任务。 开源模型代码与权重已公开发布，Studio支持按需配置多边形区域、时间跨度（1-12个月）、分辨率（10-80米）和传感器源（Sentinel-1/2）。 嵌入向量以8位有符号整数（int8）存储在COG中，提供Nano（128维）、Tiny（192维）和Base（768维）三种编码器变体。 基于冻结嵌入向量，无需重度训练即可实现相似度搜索、少样本分割（60个样本达加权F1 0.84）、变化检测与PCA无监督探索。 系统对更高性能需求支持监督微调（SFT），其表现受持续云层、大气伪影或影像缺失等输入数据质量限制。 嵌入计算与导出规格 用户可通过Studio界面或API自主指定分析区域、1至12个月的时间跨度、10至80米分辨率，并选择Sentinel-2 L2A与Sentinel-1 RTC数据源。模型变体涵盖Nano（1.4M参数/128维）、Tiny（6.2M参数/192维）和Base（89M参数/768维）。导出的数据为云优化GeoTIFF（COG），每个波段对应一个维度，数值以-127至+127的int8格式存储，支持按需量化还原为浮点数。 下游分析与典型应用 冻结嵌入支持多种快速分析：通过余弦相似度可无监督检索相似地形；利用简单逻辑回归探针，仅需60个标注像素即可在越南金瓯红树林区域实现加权F1为0.84的少样本分割；对比2023与2024年单月嵌入的余弦距离，可准确识别加州Park Fire火灾烧毁痕迹；通过PCA降维至3个通道显示为假彩色，则能自动呈现地表农田与作物结构。 微调支持与现存局限 对于精度要求更高的应用，Studio支持监督微调（SFT）训练专用任务头。官方同时提示，嵌入质量依赖于底层遥感影像质量，合成周期内的持续云层覆盖、大气伪影以及观测数据缺失都会影响最终导出的向量表现。

📄 Tech Report: https://allenai.org/papers/olmoearth | 📊 Documentation: https://docs.olmoearth.allenai.org/embeddings | 💻 Learn more about OlmoEarth: https://allenai.org/olmoearth OlmoEarth Studio , our platform for building Earth observation models, now lets you compute and export embedding vectors —compact numerical representations of Earth-observation data produced by our open source OlmoEarth foundation models. The source code and model weights are publicly available alongside the research paper , so the community can inspect exactly how these embeddings are generated. Embeddings are a fast, cost-effective entry point for leveraging OlmoEarth: they support a wide range of downstream tasks, from similarity search to segmentation to unsupervised exploration. Locations with similar surface characteristics end up with similar vectors; locations that differ land far apart. OlmoEarth embeddings have shown strong performance in our own benchmarking and in independent evaluations . The exported Cloud-Optimized GeoTIFFs (COGs) are lightweight and easy to share. Choose your area of interest, time range, encoder variant, resolution, and imagery sources via the Studio UI or API, and get back a COG you can use however you like. If your application requires higher performance, Studio also supports supervised fine-tuning (SFT) . Custom-computed embeddings are now available for users of OlmoEarth Studio. Reach out if you're interested in gaining access. Instructions for using the publicly available OlmoEarth models to compute your own embeddings are available here . Computing embeddings in Studio Global structure in OlmoEarth embeddings from seasonal Sentinel-2 imagery across 1.1M samples. Colors indicate 15 k-means clusters in a PCA-reduced embedding space. Computing embeddings follows the same workflow as any other prediction in Studio. First configure a model and run it, and then download the results. Several parameters tailor the output: Area of interest: Draw or upload any polygon; Studio handles imagery acquisition and tiling. Time span: 1-12 monthly periods. Encoder variant: Nano (128-dim, 1.4M params), Tiny (192-dim, 6.2M params), or Base (768-dim, 89M params). Spatial resolution: 10 meter, 20 meter, 40 meter, or 80 meter per pixel. Imagery sources: Sentinel-2 L2A, Sentinel-1 RTC, or both. Different visualization options applied to the same embedding raster. Studio delivers a COG with one band per embedding dimension. Vectors are stored as signed 8-bit integers (int8). Values range from -127 to +127, with -128 reserved for nodata. To recover floating-point vectors, see dequantize_embeddings in olmoearth_pretrain. Because everything is computed on demand rather than pulled from a pre-computed global archive, your embeddings reflect exactly the conditions you care about. You can generate monthly embeddings to capture seasonal dynamics, not just annual snapshots. What you can do with OlmoEarth embeddings The examples below all use OlmoEarth-v1-Tiny (192-dim) embeddings at 40-meter resolution with Sentinel-2 L2A composites (annual for most examples; monthly for change detection). Tiny is a lightweight encoder but still highly performant; for your own applications, you can swap it for a larger variant at the cost of higher compute and storage. Similarity search: Finding "more like this" Pick a query pixel, extract its embedding, and compute cosine similarity against every other pixel. The result is a heatmap showing where the landscape looks most and least like your query pixel. This query sits near the Merced urban center in California. Urban fabric and road corridors light up coherently while agricultural parcels stay dark. The model distinguishes built-up surfaces from cropland without any labels. Switching the query to a small agricultural window, we define the query vector as the mean of the embedding vectors over that window, then pull Sentinel-2 imagery at the highest- and lowest-similarity locations to see what the model treats as similar and dissimilar. The most similar patches (0.89 and above) are all agricultural parcels with irrigated fields. The least similar (around zero) are an airport with surrounding bare ground, a reservoir with dry terrain, and arid rangeland. No training data, no labels, just a dot product in embedding space. Few-shot segmentation: Labeling the landscape Similarity search tells you "where is it like this?" but sometimes you need discrete labels across a region. Because the representations are already rich, a simple linear classifier can produce a wall-to-wall land-cover map from very few labeled pixels. To test this, we labeled just 60 pixels (20 per class) over Ca Mau, Vietnam, a coastal mangrove region. Using ESA WorldCover 2021 as the label source for three classes (mangrove, water, other), we randomly sampled 20 pixels per class, trained a logistic regression with per-feature standardization, and predicted every pixel in the region. From 60 labeled pixels, the classifier produces a coherent map with weighted F1 = 0.84 . Mangrove stands, tidal channels, and open water are delineated across the entire region. The classifier saturates quickly: increasing from 30 to 300 labels barely changes accuracy, because the embeddings are doing most of the heavy lifting. The core of the analysis is a few lines of Python: import rasterio import numpy as np from sklearn.pipeline import make_pipeline from sklearn.preprocessing import StandardScaler from sklearn.linear_model import LogisticRegression # Load the 192-band embedding COG exported from Studio with rasterio. open ( "embeddings.tif" ) as ds: emb = ds.read().astype(np.float32) # (192, H, W) C, H, W = emb.shape X = emb.reshape(C, - 1 ).T # (H*W, 192) # Train on labeled pixels, predict everywhere clf = make_pipeline(StandardScaler(), LogisticRegression(max_iter= 2000 )) clf.fit(X[train_idx], labels[train_idx]) prediction = clf.predict(X).reshape(H, W) This is a linear probe , a standard evaluation for foundation models. The fact that a logistic regression over 192 dimensions recovers land-cover boundaries from so few labels means the Tiny encoder has organized these ecological distinctions during pretraining. Larger variants (Base, 768-dim) encode even richer representations. If you have ground-truth polygons, field survey points, or a coarse existing map, you can train a similar classifier and produce a wall-to-wall map for your own region of interest. Change detection: Spotting what shifted Because Studio can generate embeddings at any temporal resolution (monthly through annual), you can compare two time periods directly to identify where surface conditions have changed. Below, we computed monthly Sentinel-2 embeddings for the same region in September 2023 and September 2024 and measured per-pixel cosine distance. The Park Fire (July-September 2024) burn scar in Butte County, California lights up immediately. No labels or training required—just two embedding COGs and a few lines of Python. Unsupervised exploration: Seeing what the model sees Sometimes you have no query location or reference labels. You just want to understand what structure exists in the embeddings. Principal Component Analysis (PCA) is a clean way to do this: reduce to three dimensions, map to R/G/B, and display as a false-color image. Similar embeddings get similar colors automatically. Flevoland, in the Netherlands, is a reclaimed polder landscape with a regular grid of agricultural parcels. The PCA false-color image reproduces those boundaries with high fidelity. Different crop types, water bodies, and urban areas each get distinct hues. The embedding has internalized landscape structure without ever being told what a parcel or crop is. This kind of unsupervised view is a quick way to see what structure the model has picked up across your area of interest. From export to insight Similarity search, few-shot segmentation, change detection, and PCA exploration are simple operations on standard raster data that run in seconds. The power comes from the embeddings: learned representations that compress earth observation data into vectors capturing rich information about each location from many sensors and millions of training examples. Custom embedding exports are available now. Create a project, configure an embeddings model, and compute your embeddings . The exported GeoTIFF works with any geospatial tool: QGIS, GDAL, rasterio, or your own scripts. For end-to-end code reproducing the examples in this post, see the embeddings tutorial , which includes working code for similarity search , few-shot segmentation , change detection , and PCA visualization . To get hands-on without any local setup, try the Colab notebook . Going further: fine-tuning The examples in this post all use frozen embeddings with no task-specific training. Embeddings are a great entry point for leveraging OlmoEarth: they enable fast, cost-effective generation of results, work well in resource-constrained environments, and are easy to share. For applications that require higher performance, OlmoEarth Studio also supports SFT , training a task-specific model head on your own labels, which typically outperforms linear probes on frozen features. Limitations While we are always working to improve our pretraining approaches, it's important to check the quality of the embeddings for your use case using some of the techniques described above. Performance also depends on the quality of the input imagery—persistent cloud cover, atmospheric artifacts, or missing observations in the composite period can affect the resulting vectors. Sentinel-2 L2A imagery from the European Space Agency, accessed via Microsoft Planetary Computer. ESA WorldCover 2021 v200 for mangrove reference. The global clustering visualization uses OlmoEarth-v1-Base (768-dim) with seasonal Sentinel-2 composites at patch size 8, instance-level embeddings over 1.1M pretraining samples.