DanDan1505/-PROSPEKT

PROSPEKT — An AI-powered mineral prospectivity mapping system for greenfield geological exploration. Uses Sentinel-2 satellite imagery, Random Forest & XGBoost classifiers, and a Flask dashboard to predict and visualise mineral-rich zones. Built for mining operations in Ghana and the developing world. SDG 8 · SDG 9 # PROSPEKT AI-powered mineral prospectivity mapping for greenfield geological exploration in Ghana. PROSPEKT uses free satellite and elevation data to identify zones that may be worth field investigation. The current prototype focuses on the Tarkwa gold belt area and includes a machine learning pipeline plus an offline-friendly Flask dashboard. ## What It Does ```text Satellite + elevation data -> geospatial feature extraction -> Random Forest and XGBoost models -> prospectivity scores from 0-100% -> interactive dashboard map ``` Important: the current labels are weak labels from remote-sensing heuristics, not confirmed mineral deposits. The system prioritizes zones for investigation; it does not prove minerals are present. ## Current Status ```text Phase 1: Data pipeline + ML model Complete Phase 2: Flask dashboard Complete Phase 3: Arduino sensor node Next ``` ## Main Features Built - Downloads Sentinel-2 and SRTM data for Tarkwa, Ghana. - Extracts spectral, vegetation, alteration, elevation, and slope features. - Trains Random Forest and XGBoost classifiers. - Produces CSV, GeoJSON, and GeoTIFF prospectivity outputs. - Shows prospectivity zones on a Leaflet dashboard map. - Supports plain-language filtering such as `high iron oxide near Tarkwa`. - Flags zones above a confidence threshold. - Works offline for local zone data, filters, alerts, and dashboard controls. ## Quick Start Create and activate the environment: ```powershell python -m venv .venv .\.venv\Scripts\Activate.ps1 pip install -r requirements.txt ``` Run the dashboard: ```powershell python dashboard\run.py ``` Open: ```text 127.0.0.1 ``` Keep the terminal open while using the dashboard. ## Key Outputs ```text outputs/scores/tarkwa_final_zone_scores.csv outputs/maps/tarkwa_final_zone_scores.geojson outputs/maps/tarkwa_random_forest_prospectivity.tif outputs/figures/feature_importance_comparison.png dashboard/prospekt.db ``` ## Model Result …