Temporal Embeddings of Surface Spectra for Earth Representation and Analysis (TESSERA) [CVPR2026]

<div align="center"> <a href="#readme-top"> <img src="images/banner.png" alt="Banner"> </a> <br /> <p align="center"> <a href="https://geotessera.org/">Project Website 🌐</a> &nbsp;&nbsp;&nbsp;&nbsp; <a href="https://github.com/FrankFeng-23/btfm_project/issues">Report Bugs 🛠️</a> &nbsp;&nbsp;&nbsp;&nbsp; <a href="https://github.com/FrankFeng-23/btfm_project/issues">Request Features 💡</a> </p> </div> <!--  -->

Table of Contents

• Learning about TESSERA
  • Introduction
  • Papers
  • Podcast
  • Presentations
  • License
• Using TESSERA
  • Acceptable Use Policy
  • Accessing Precomputed Embeddings
  • Creating Your Own Embeddings
  • Downstream Tasks
  • TESSERA Users Group
• Additional information
  • Team
  • Contact
  • Citation
  • Acknowledgments
  • Star History

Learning about TESSERA

Introduction

Satellite remote sensing enables a wide range of downstream applications, including habitat mapping, carbon accounting, and strategies for conservation and sustainable land use. However, satellite time series are voluminous and often cloud-corrupted, making them challenging to use: the scientific community's ability to extract actionable insights is often constrained by the scarcity of labelled training datasets and the computational burden of processing temporal data. The key insight behind our work, due to Dr. Clement Atzberger is that forcing auto-encoder embeddings derived from two cloud-free random samples of satellite time series to align using Barlow Twins results in an embedding that represents the entire time series, including the missing observations.

This idea is the key behind TESSERA, an open foundation model that preserves per-pixel spectral-temporal signals in 128-dimensional latent representations at 10-meter resolution globally. It uses self-supervised learning to summarise petabytes of Earth observation data. We compare our work with state-of-the-art task-specific models and other foundation models in five diverse downstream tasks and find that TESSERA closely matches or outperforms these baselines. By preserving temporal phenological signals that are typically lost in conventional approaches, TESSERA enables new insights into ecosystem dynamics, agricultural food systems, and environmental change detection. Moreover, our open-source implementation supports reproducibility and extensibility, while the privacy-preserving design allows researchers to maintain data sovereignty.

To our knowledge, TESSERA is unprecedented in its ease of use, scale, and accuracy: no other foundation model provides analysis-ready outputs, is open, and provides global, annual coverage at 10m resolution using only spectral-temporal features at pixel level.

Here are some visualization results of the TESSERA representation map (using the first three channels as RGB):

Papers

Here are publications and preprints related to TESSERA, listed chronologically:

• Lisaius, M. C., Blake, A., Keshav, S., & Atzberger, C. (2024). Using Barlow Twins to Create Representations From Cloud-Corrupted Remote Sensing Time Series. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 17, 13162–13168. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. https://doi.org/10.1109/JSTARS.2024.3426044

• Z. Feng, C. Atzberger, S. Jaffer, J. Knezevic, S. Sormunen, R. Young, M.C. Lisaius, M. Immitzer, T. Jackson, J. Ball, D.A. Coomes, A. Madhavapeddy, A. Blake, S. Keshav (2025), TESSERA: Temporal Embeddings of Surface Spectra for Earth Representation and Analysis, To Appear, CVPR 2026. ArXiv reprint. https://arxiv.org/abs/2506.20380

• Lisaius, M. C., Blake, A., Atzberger, C., & Keshav, S. (2026). Towards improved crop type classification: A compact embedding approach suitable for small fields. Accepted in Proceedings of the ISPRS Conference 2026. International Society for Photogrammetry and Remote Sensing.

• Z. Feng, C. Atzberger, S. Jaffer, J. Knezevic, S. Sormunen, R. Young, M.C. Lisaius, M. Immitzer, T. Jackson, J. Ball, D.A. Coomes, A. Madhavapeddy, A. Blake, S. Keshav, (2026) Applications of the TESSERA Geospatial Foundation Model to Diverse Environmental Mapping Tasks, SSRN preprint. http://ssrn.com/abstract=6142416

• Young, R., & Keshav, S. (2026). Interpolation of GEDI Biomass Estimates with Calibrated Uncertainty Quantification, arXiv preprint. https://doi.org/10.48550/ArXiv.2601.16834

• Lisaius, M. C., Keshav, S., Blake, A., & Atzberger, C. (2026). Embedding-based Crop Type Classification in the Groundnut Basin of Senegal (arXiv:2601.16900). ArXiv preprint. https://doi.org/10.48550/arXiv.2601.16900

• Ball, J.G.C, Wicklein J.A. , Feng, Z., Knezevic, J., Jaffer, S., Atzberger, C., Dalponte, M., and Coomes, D. Geospatial foundation models enable data-efficient tree species mapping in temperate montane forests, BioArxiv, https://doi.org/10.64898/2026.02.23.707022

Presentations

• TESSERA overview in AI for Good seminar, Frank Feng, Jan 22, 2026
• TESSERA: Precomputed FAIR Global Pixel Embeddings for Earth Representation and Analysis IEEE GRSS Talk, Frank Feng, 12 December, 2025
• 2-slide summary (PPTX) for CRI Flash Talks, S. Keshav, October 7, 2025
• Foundation model overview (PPTX) for Ecology Groups meeting, University of Cambridge, DAB, James Ball, October 6, 2025
• TESSERA overview presentation with a focus on ecological applications (PDF) University of Maryland, Frank Feng, October 1, 2025
• TESSERA overview presentation (PPTX) James Cook University, S. Keshav, September 29, 2025
• TESSERA overview presentation University of Cambridge, DAB, Frank Feng, May 20, 2025
• Self-supervised learning for earth observation (PPTX) S. Keshav, Exeter, April 2025

License

TESSERA software is released under the standard MIT license. Embeddings and model weights are released under the CC0 license: essentially, they can be freely used for both commercial and non-commercial purposes. Although we do not legally require attribution, we do request it.

Using TESSERA

<a id="global-embeddings-access"></a>

Accessing Embeddings using GeoTessera (recommended)

We have generated embeddings for the whole globe at 10m resolution for 2024. These can be downloaded and used for downstream applications, saving significant computational time and resources, using the GeoTessera library. We will progressively extending coverage backwards year by year until 2017. The current coverage map is below:

<img src="https://github.com/ucam-eo/tessera-coverage-map/blob/main/map.png">

TESSERA Users Group

Interested users are invited to join our Zulip discussion groups.

Creating Your Own Embeddings

If you would like to use our software to create your own embeddings, please follow the instructions below. Note that this is a comptuationally challenging task and you will need access to significant computational and storage resources.

Hardware Requirements

1. Storage Requirements

Running this pipeline requires substantial storage space. Although the pipeline cleans up some intermediate files after processing, the downloaded raw Sentinel-2 and Sentinel-1 files will still occupy considerable disk space. For example, processing a 100km×100km area from 2022 to output a TESSERA Representation map (10m resolution) requires at least 1TB of storage.

2. Memory Requirements

We use preprocessed data, initially from Microsoft Planetary Computer. However, the next generation of embeddings will use OPERA from ASF DAAC. In either case, most of the geo-preprocessing has been done. Still, we recommend having at least 128GB of RAM.

3. CPU and GPU

The pipeline has no strict requirements for CPU and GPU, but more CPU cores and more powerful GPUs can significantly speed up inference. When processing a 110km×110km area from 2022, our tests using a 128-core CPU and a single NVIDIA A30 GPU for inference (CPU and GPU each handling 50% of the inference) took approximately 10 hours to complete.

4. Operating System

For the data preprocessing pipeline, we support almost all Linux systems. For Window