Mining · Earth observation
ZAHIR — AI satellite mineral exploration
From pixel to target: turns satellite imagery and elevation models into prioritized mineral exploration targets. Under active development; field calibration and validation depend on geological context.
Sentinel-2ASTEREMITIA
- Problem
- Early mineral exploration is expensive, slow and scattered: large areas and heterogeneous data (spectrum, topography, regional geology) for campaign decisions worth millions.
- Solution
- ZAHIR turns satellite imagery and terrain elevation models into prioritized exploration targets: spectral mineral-alteration detection, terrain structural analysis —faults, gullies and their intersections— and a fusion of both with a priority score. It tells the geologist where to look first.
- Technology
- Sentinel-2, ASTER and the EMIT hyperspectral sensor; AI and computer vision; metallogenic context of Andean porphyry belts. Web geospatial viewer, REST API and GeoJSON export ready for QGIS and ArcGIS.
- Outcome
- Highest-potential zones distilled automatically, shrinking the area to survey on the ground before moving a single crew. Reproducible, explainable methods: every result traces back to its scene and parameters. The platform is under active development; fine calibration and field validation adapt to each geological context.