Guide · 8 min read
What is stockpile volume measurement, and how accurate can it actually get?
Industrial operators have measured stockpiles with rope, tape, profile boards, and expert eye for the better part of a century. Drone photogrammetry changed the economics of doing it right. This is a working guide to what the method does, where the accuracy actually comes from, and where it breaks down.
The short answer
A drone-based stockpile measurement, properly executed, delivers volume accuracy in the ±1.5% range against weighbridge-reconciled ground truth. That compares to 5–10% variance that most manual-survey regimes carry. At a coal yard moving half a million tonnes a cycle, that gap is not an engineering number — it is a balance-sheet number.
How the method works
A drone flies a pre-planned grid over the stockpile at a fixed altitude — typically 40–80 metres above the highest point of the pile. A camera (RGB, 24+ megapixels, fixed focal length) captures 250–600 images with 75% forward overlap and 65% side overlap. Ground control points (GCPs) — surveyed reference markers — are placed around the perimeter. Processing software (photogrammetric bundle adjustment) turns the image set into a dense point cloud, then into a digital surface model (DSM). Volume is the difference between the DSM and a defined base surface (toe polygon).
LiDAR is used where vegetation, tarpaulin, or low-contrast surfaces would defeat photogrammetry — typically outdoor overburden dumps and iron ore concentrate yards during monsoon.
Where the accuracy actually comes from
Three factors dominate. GCP density and survey quality — you cannot out-process a badly-surveyed GCP. Flight altitude and image overlap — too high, and your ground sample distance (GSD) exceeds the tolerance budget; too low, and you lose overlap. Base-surface definition — where is "zero"? The toe polygon must be defined consistently cycle-over-cycle. Changing it between measurements introduces phantom volume.
Why weighbridge reconciliation is non-negotiable
Volume is the measurement; tonnage is the number the business runs on. Converting one to the other requires a bulk-density assumption. That assumption is not constant — it varies with moisture, particle size, compaction, and time since tip. A measurement programme that does not reconcile against weighbridge throughput is a measurement programme that is silently wrong.
Flybi's delivery always includes a reconciliation table: drone-derived tonnage against weighbridge inflow/outflow for the measurement period, with the residual flagged if it exceeds tolerance. That residual is either a measurement error (investigate) or a real inventory discrepancy (act).
Where the method breaks down
Four failure modes. Live conveyors and active handling — you cannot measure a surface that is moving. Schedule the flight in a standstill window. Waterlogged ground — reflective water defeats photogrammetry; use LiDAR or wait. Tarpaulin cover — measures the tarp, not the pile; coordinate removal. Thermal plumes — over hot coal or clinker, air turbulence distorts image geometry; fly higher or at dawn.
The CFO test
Ask your current stockpile auditor four questions. What is your reported accuracy envelope, referenced to what? How do you reconcile to the weighbridge? What is your base-surface definition protocol? What is your re-survey cadence? If the answers are not crisp, the closing inventory number is a negotiation, not a measurement.
Related reading
From reading to doing
Bring us a real site, a real block, a real question.
Most of our engagements start here — a reader who became a customer inside one scoping call.