Tessera AI mannequin affords accessible approach to view Earth

Tessera, a sophisticated synthetic intelligence (AI) mannequin, affords high-accuracy datasets that encode what the satellite tv for pc ‘sees’ of Earth’s floor throughout the course of a yr. This compressed knowledge can be utilized by the scientific group to generate information-rich maps.

Crucially, the encoded datasets – known as ’embeddings’ – use far much less knowledge than the pixellated photographs which might be transmitted to Earth from satellites. A wide range of purposes are supported by the mannequin, from monitoring agricultural crops to measuring areas burnt by fireplace and forest canopies.

A paper on Tessera was revealed on the 2026 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), held 3-7 June. The mannequin itself was first launched in 2025 and the paper marks the first fully peer-reviewed announcement of Tessera to the scientific group.

The inspiration mannequin – Temporal Embeddings of Floor Spectra for Earth Illustration and Evaluation, or Tessera for brief – was developed by researchers on the College of Cambridge within the UK, alongside world and European companions, together with Aalto College in Finland.

Tessera’s processed datasets, or embeddings, provide a number of particular advantages to the Earth commentary group. As a result of Tessera’s embeddings are pretrained, they seize patterns within the knowledge and modifications over time that different strategies should study from scratch. Which means non-AI consultants can clear up distant sensing issues at a worldwide scale utilizing solely a fraction of the labelled knowledge beforehand required. The embeddings are additionally light-weight sufficient to entry from a laptop computer or perhaps a cellular gadget, making them out there to customers with out computational sources. And as an open-source mission, it’s freely modifiable, elevating close to limitless prospects for utilizing satellite tv for pc datasets to review the Earth.

In line with Nuno Miranda, Mission Supervisor for Sentinel-1 on the European Area Company (ESA), that is an revolutionary and thrilling step within the improvement and use of AI within the area of Earth commentary. He stated, “Basis fashions are the brand new frontier of AI utilized to remote-sensing knowledge. Tessera demonstrates how knowledge from the Sentinel-1 and Sentinel-2 missions may be utilized in follow, serving to customers to analyse and perceive the Earth system extra effectively.”

Srinivasan Keshav, professor on the College of Cambridge and co-lead on the Tessera mission, famous, “With Tessera, we have addressed a number of the challenges of working with the very giant quantities of information supplied by the Copernicus programme. Our embeddings make the info extra accessible to customers from historically unserved communities, particularly these from ecology, conservation, plant science and zoology. We’ve additionally made these out there with out requiring registration and without charge, opening the door to many new courses of essential issues.”

What’s Tessera?

Tessera processes enormous quantities of remote-sensing knowledge from the Copernicus missions, Sentinel-1 and Sentinel-2. It combines two types of data: optical data from Sentinel-2, and advanced radar data, known as synthetic aperture radar (SAR) data, from Sentinel-1. The optical and radar datasets are fused by the foundation model and processed into global embeddings spanning each year from 2017-2025.

So, rather than the data-heavy and pixellated imagery from satellites, Tessera compresses data heavy, cloudy satellite imagery to create an embedding layer of Earth data. It does this at a resolution of 10 m, which is the same as the highest resolution captured by Sentinel-2.

Tessera’s embedding layers are basically compressed Earth observation data with missing values filled in. Each 10-m pixel contains a time series of what happened at that point over the year. This gives researchers a picture of change – rather than how a field, river or mountain looks at any given point in time – in a format that’s searchable.

Tessera is supported by tools that enable users to search and compare Earth imagery in a number of different ways. For example, users can search for geographic regions that are similar to each other and they can look for changes in landscapes. It is also possible to make predictions about vegetation health and urban growth.

Tracking habitat change

A UK-based project involving Tessera is developing ways to evaluate the UK government’s nature protection schemes using satellite data from Sentinels 1 and 2. Researchers used Tessera embeddings to track habitat change on land designated for protection across Cumbria, an area of northern England. The project, a partnership between Tessera, the Endangered Landscapes and Seascape Programme, and other UK partners, could eventually provide the government with a way to measure the effectiveness of investments in farming subsidies and nature conservation.

One of Tessera’s co-leads and a senior researcher on the Cumbria landscape monitoring project, Professor David Coomes, from the University of Cambridge, said, “Monitoring these environmental changes over vast scales is exactly the sort of problem that Tessera was designed to solve.”

How are foundation models changing Earth observation?

Tessera promotes transparency and reproducibility. It is open-source and aligned with FAIR principles (Findable, Accessible, Interoperable, Reusable) – a set of widely adopted guidelines developed by the international research community on the reusability of digital assets.

It offers an open and transparent alternative to systems such as AlphaEarth Foundations, an AI model by Google DeepMind, which also compresses complex satellite data from multiple sources to create embeddings using a closed model.

Moreover, Tessera facilitates access to Copernicus data and offers an efficient way to explore Earth observation data.

According to the University of Cambridge’s Srinivasan Keshav, “The adoption of embeddings represents a paradigm shift. Instead of distributing heavy imagery, models such as Tessera can now provide downstream users with compressed semantic representations of the information of Earth’s surface embedded in the original data.”

ESA partners on models for Earth observation

Several teams are working on foundation models for Earth observation, placing Europe at the forefront of this field. ESA has also pioneered the development of foundation models trained on Earth observation data through its open innovation laboratory, Φ-lab, a hub and catalyst for Earth observation innovation. Two foundation models to recently come out of Φ-lab are Thor, developed by the Norwegian Computing Centre, and TerraMind, developed with IBM Research Europe. Importantly, unlike models such as Tessera or AlphaEarth that aggregate information into long-term or annual embeddings, both Thor and TerraMind focus on learning from individual observations, preserving rich spatial contextual information within single imagery snapshots rather than encapsulating everything into a single compressed representation.

Thor (Transformer-based basis mannequin for Heterogeneous Commentary and Decision) is a flexible multi-modal basis mannequin, which mixes several types of knowledge and is designed to beat each the challenges of assorted inputs and inflexible deployment constraints. Whereas most present basis fashions are architecturally inflexible, Thor permits the person to adapt the inner decision and optimise computational efficiency. It’s educated on knowledge from Sentinels 1, 2 and three. This mannequin was funded and supported by ESA’s Basis Fashions for Local weather and Society (FM4CS) mission.

TerraMind, a basis mannequin launched in April 2025, can be multi-modal and capable of reply questions on local weather and nature. Somewhat than focusing solely on downstream duties, its core innovation lies in studying a unified illustration area that aligns a number of sources of geospatial knowledge – together with satellite tv for pc imagery, topography, land use/land cowl, elevation and geolocation. This permits cross-modal reasoning and query-based interplay with Earth system knowledge. By collectively embedding these various sources, TerraMind strikes past conventional task-specific fashions towards a extra general-purpose geospatial intelligence framework. It was educated on a dataset of greater than 9 million globally distributed samples spanning eight complementary knowledge varieties, together with radar from Copernicus Sentinel-1 and optical Sentinel-2 imagery.