Why drones replaced GPS-rover stockpile surveys
Three operational facts drive the shift in the Saudi mining sector:
- Speed. A GPS-rover crew takes 4–8 hours to survey a 100,000 m3 stockpile. A drone takes 25–45 minutes for the same pile [VERIFY-SME].
- Density. A rover survey produces a few hundred ground points; a drone produces millions. The DEM resolution is an order of magnitude tighter.
- Defensibility. The drone deliverable is a georeferenced point cloud the auditor can re-process; the rover deliverable is a spreadsheet.
Phosphate, gold and bauxite operations across Ma’aden assets have moved to drone-first methodology by 2026 [VERIFY-SME]. The remaining rover-only surveys are typically on stockpiles too small to justify mobilisation.
Three accuracy bands you should know
| Band | Volume RMSE | Required for |
|---|---|---|
| Reconnaissance | ±5% | Operational planning |
| Operational | ±2.5% | Monthly internal close |
| Audit-grade | ±1.5% | External financial audit |
A 2026 Matrice + RTK + dense photogrammetry hits the operational band consistently and the audit-grade band on most stockpile geometries. Lidar is required only for unusually steep or vegetated piles — see the photogrammetry vs lidar piece.
Methodology pack — what auditors expect
A defensible methodology pack contains:
- RTK base configuration — base location, antenna height, observation duration, NRTK or PPK selection.
- GCP layout — at least 5 GCPs distributed around and on top of the pile, surveyed against the Saudi National Address datum or the operating organisation’s geodetic frame.
- Flight plan — overlap (typically 80% forward, 70% side), altitude, GSD target, weather window.
- Point cloud processing — software version, filtering parameters, ground classification logic.
- Volume computation — base surface definition (toe of pile, planar, design), buffer zone, calculation tool.
- Chain of custody — raw imagery hash, processed point cloud hash, exported volume report signed by the surveyor and the operations engineer.
Anchor the workflow in the drone site survey solution and the 3D site mapping solution.
RTK setup that survives audit
Three rules that separate audit-grade from operational-grade work:
- Base on a known monument, not a tripod. Tie the survey to the operating organisation’s existing geodetic network.
- Observation duration ≥ 1 hour at the base before flight, with documented horizontal and vertical RMSE.
- PPK as a fallback when NRTK loses fix. PPK adds ~30 minutes of post-processing but salvages otherwise unusable flights.
For the underlying primitives see the RTK GPS glossary and the ground control point glossary.
The base surface decision
Volume = (top surface) − (base surface). The top surface comes from the drone. The base surface is a methodology choice that auditors will scrutinise:
- Toe-of-pile: surface defined by the survey crew tracing the pile foot. Most accurate but most labour-intensive.
- Planar: a best-fit plane through the surrounding pad. Fast, but inaccurate on uneven yard floors.
- Design surface: a CAD/BIM-defined yard floor. Good for greenfield yards, poor for established ones with subsidence.
- Previous survey: differential volume against a prior point cloud. Excellent for change-detection, but needs the prior survey to have been audit-grade.
Pick one, document it, freeze it for the year. Switching mid-year is an audit red flag.
False positives — what produces volume errors
Five recurring sources of error on KSA stockpiles:
- Wind-blown dust in the upper image stack — adds spurious height.
- Heavy machinery on the pile during flight — doubles as height.
- Shadow noise at low solar elevation (early morning, late afternoon) — distorts photogrammetry depth.
- Reflective wet surfaces after rain — confuses dense matching.
- Sand encroachment on the pad — shifts the base surface.
A 2026 methodology pack documents how each is mitigated.
Integration with the monthly close
The financial close cycle has hard dates. The drone team has weather. Three patterns work:
- Two-week survey window at month-end with a 5-day weather buffer.
- Differential surveys from a known audit-grade baseline; each month is a delta, not a full survey.
- Triple-redundancy — three flights across the window, with the median volume reported.
Tie the deliverable to the progress tracking solution and the AI analytics platform for change-over-time visualisation.
Cost envelope
Indicative SAR cost ranges for 2026 KSA mining stockpile work:
| Pile size | One-off survey | Annual programme (12 surveys) |
|---|---|---|
| < 50,000 m3 | 8,000–14,000 | 70,000–110,000 |
| 50,000–200,000 m3 | 14,000–28,000 | 130,000–220,000 |
| 200,000–1M m3 | 28,000–55,000 | 250,000–450,000 |
| > 1M m3 | 55,000–95,000 | 480,000–800,000 |
[VERIFY-SME for KSA-specific pricing.] See the drone mapping cost piece for the broader pricing context.
Common mistakes that fail the audit
- Single GCP or no GCPs — the auditor cannot validate the geodetic anchor.
- Rolling base surface — switching definitions mid-year.
- No point-cloud archive — the auditor cannot re-process.
- Mixed crews — different surveyors using different software versions across the year.
Field deployment checklist
- Methodology pack approved by finance and audit.
- RTK base on a known monument with ≥1 hour observation.
- GCP layout signed by the surveyor.
- Base-surface decision frozen for the year.
- Chain-of-custody hashes recorded.
- Two months of parallel rover-vs-drone validation before cut-over.
Next steps
If you are scoping volumetric drone surveys in the Saudi mining sector, start with the drone site survey solution, the 3D site mapping solution and the drone mapping cost piece. Cross-reference the photogrammetry vs lidar piece.
Book a stockpile-survey scoping call and we will produce a methodology pack and SAR cost envelope for your specific operation within 10 working days.
