What the $1M will fund
This is not vague AI funding. It goes into concrete infrastructure that improves speed, precision, and credibility of galamsey detection.
- Visibility: satellite coverage and refresh frequency
- Processing: compute for fine-grid and high-volume analysis
- Intelligence: model training and Ghana-specific tuning
- Action: alerts, dashboard workflows, and validation
Detailed Allocation
Satellite Data
$250K – $350K- High-resolution imagery (sub-5m where needed)
- Frequent refresh cycles (not occasional snapshots)
- Optical + SAR (radar) coverage
This enables earlier detection, not just after visible damage.
Compute Infrastructure
$300K – $400K- GPU servers for training detection models
- Large-scale processing across billions of grid cells
- Cloud + hybrid compute for continuous scanning
This powers a 10m × 10m grid strategy (~2.38B cells).
AI Model Development
$150K – $200K- Segmentation model training for mining signatures
- Ghana-specific model tuning and calibration
- Continuous updates as patterns shift
Detection becomes intelligent, not only visual.
Data Pipeline & Platform
$100K – $150K- Satellite feed processing into usable intelligence
- Real-time alerting and triage workflows
- Monitoring dashboard for decisions and response
Turns raw imagery into actionable operations.
Field Validation & Data Collection
$50K – $100K- Ground-truth verification
- Drone validation in high-risk zones
- Ghana-specific labeled dataset expansion
Improves confidence, credibility, and precision.
Current Position and Upgrade
Current system (honest baseline)
Today the scan grid is medium-to-coarse, roughly in the 250m–1km range.
- 1km × 1km cell = 1,000,000 m²
- 250m × 250m cell = 62,500 m²
- Mixed land signals in one cell reduce precision and delay early warnings
Upgrade target
10m × 10m cells (100 m²), estimated ~2.38 billion cells across Ghana.
- Detect smaller disturbance earlier
- Improve location precision for response teams
- Reduce false positives through finer context