GeoMiner: Mineral Exploration Targeting System
Problem
A single 500m diamond drill hole costs ~$250,000. Over exploration license blocks spanning 50 to 5,000 km², blind drilling is financially prohibitive. Early-stage exploration teams require rapid, regional generative reconnaissance to eliminate barren terrain and pinpoint high-probability alteration zones before deploying field crews.
Approach
Architected a full-stack, cloud-based prospectivity targeting platform powered by Google Earth Engine, Sentinel-2 L2A multispectral imagery, and Copernicus DEM. Developed custom band ratio pipelines for Hydroxyl/Clay (SWIR1/SWIR2), Ferric Iron (Red/Green), and Gossan (SWIR1/Red) with automated NDVI vegetation masking. Integrated multi-directional shaded-relief convolutions to detect structural fault scarps and lineament intersection corridors. Implemented Analytic Hierarchy Process (AHP) multi-criteria decision matrices across 5 genetic mineral deposit models (Porphyry Cu-Au-Mo, Epithermal Au-Ag, Lithium LCT, Iron Oxide, and VMS) with automated active mine disturbance suppression.
Tech Stack
Results
Compresses months of regional manual GIS preprocessing into on-demand cloud analysis running in minutes. Generates ranked vector polygon targets with surface footprints (ha / km²), centroid coordinates, and relative prospectivity ranking scores, with one-click export to GeoJSON, CSV summary, 3D DXF (CAD), and Leapfrog Geo drill hole collars.
What I Learned
Mastered multi-criteria spatial decision analysis (AHP) combining raster spectral alteration with vector geophysical structures; solved critical edge cases in satellite optical remote sensing including cloud masking, terrain shadows, and distinguishing natural bedrock anomalies from active open-pit mine infrastructure.
Enjoyed this piece?
Leave a reaction to let the author know what resonated with you.