EXPLOMAT

Sun, 01 Jun 2025 00:00:00 +0000

EXPLOMAT (Sustainable Exploration Technologies for Efficient Raw Materials Discovery) is an innovation project awarded by EIT RawMaterials under Horizon Europe, in its call for regional innovation. It runs for two years, from June 2025 to May 2027. TerraEye is the lead partner: it develops the technology, manages the project and prepares the product for the market.

The aim is a set of AI tools, built on Earth Observation data, that make mineral exploration faster, cheaper and less invasive than manual field surveys and interpretation that depends on the expert. TerraEye works on three components: automatic detection of mineral outcrops in multispectral satellite images, a model that finds stress in vegetation caused by metal-rich soils, which allows exploring forested areas without clearing them, and semi-automated analysis of geological structures from satellite and terrain data.

In the second phase of the project the tools are to be validated in a working bauxite mine in Bosnia and Herzegovina, which is the first test of the product in an industrial setting.

RISPECT

Sat, 01 Feb 2025 00:00:00 +0000

RISPECT (Regional Innovation Scheme Project for Exploration Capabilities and Technologies) is a research and innovation project awarded by EIT RawMaterials under Horizon Europe, in its programme for regional innovation. It runs from February 2025 to the end of 2026 and aims to introduce and validate technologies for exploring critical raw materials, such as lithium, bauxite, copper and nickel, in regions of Poland, Greece, Albania and Ukraine, so that exploration costs less and puts less strain on the environment.

Three technologies come together. TerraEye leads the work on synthetic hyperspectral imagery generated from multispectral satellite data, WIDMO from Poland provides a ground-penetrating radar system, and SAISLAB from Greece an AI platform that brings the data of both together.

What we want to achieve. A hyperspectral sensor records hundreds of narrow bands, so every mineral shows up with its own spectral signature, but such data are scarce and cover small areas. Multispectral satellites such as Sentinel-2 and ASTER have only a few broad bands, but they cover the whole world. We are not trying to reproduce the hyperspectral signature itself. What we want to reproduce are the analytical capabilities that come from it, such as recognising minerals and groups of minerals, and to do that by fusing the data of the multispectral sensors.

What makes it difficult. The sensors have different spatial resolutions, from 10 to 30 metres, and their data come from different periods: Sentinel-2 is a current source, while the shortwave infrared archive of ASTER is historical. The data cannot simply be stacked on top of each other. The method has to combine the complementary information of both sensors in one common representation, to say how certain its answer is, and to be clear about what cannot be recovered. The result is a spectral representation aligned with hyperspectral data, not a claim to recover information that is missing from the input.

Projects |

EXPLOMAT

RISPECT