Five years ago, running a drone survey over an open-pit mine was a novelty. Today, it’s difficult to find a mine site in Australia or South America that doesn’t use drones for at least some portion of its survey work. The shift has been remarkably fast, driven by obvious advantages in speed, safety, and cost.

But the conversation around drone surveying has matured past the initial enthusiasm. The question is no longer “should we use drones?” but rather “how accurate are they, where do they fall short, and what are we missing when we rely on them exclusively?”

What Drones Do Well

The fundamental advantage of drone surveying in an open-pit environment is coverage. A traditional ground survey using a total station or GNSS rover might capture a few hundred points across a pit face in a day. A drone carrying a high-resolution camera or LiDAR sensor can capture millions of points across the entire pit in a single flight lasting 30-60 minutes.

This density of data means that stockpile volume calculations, pit wall stability assessments, and progress monitoring can be performed far more frequently and comprehensively than was possible with traditional methods.

DJI’s Matrice series and the Wingtra fixed-wing platforms have become the workhorses of mining survey operations. Photogrammetry software from Pix4D and Agisoft processes the raw imagery into georeferenced point clouds and orthomosaics that can be ingested into mine planning software.

The safety benefit is genuine and significant. Surveying active pit faces, unstable walls, and highwall areas previously required surveyors to work in environments with real geotechnical risk. Drones remove the person from the hazard while providing better data than manual methods could achieve.

Where Accuracy Gets Complicated

Vendor marketing often claims survey-grade accuracy from drone platforms. “Sub-centimetre accuracy” appears in brochures from nearly every drone survey provider. These claims aren’t exactly wrong, but they’re not exactly right either.

Photogrammetric accuracy depends on multiple variables: ground sample distance (which is a function of flight altitude and camera resolution), ground control point density and distribution, surface texture, lighting conditions, and processing parameters. Under ideal conditions — good GCP coverage, textured surfaces, consistent lighting — photogrammetric surveys from drones can achieve horizontal accuracy of 2-3cm and vertical accuracy of 3-5cm.

But ideal conditions don’t always exist on mine sites. Several factors degrade accuracy in practice.

Smooth or Reflective Surfaces

Photogrammetry relies on identifying matching features across overlapping images. Surfaces that are smooth, wet, or uniformly coloured — like fresh bench faces in light-coloured limestone or wet clay surfaces after rain — lack the texture needed for reliable feature matching. Accuracy degrades substantially on these surfaces, sometimes to 10-20cm or worse.

LiDAR-equipped drones mitigate this issue because LiDAR doesn’t depend on visual texture. However, LiDAR drone payloads are heavier, more expensive, and typically have shorter flight times.

Vegetation and Overburden

In mines where vegetation encroaches on bench edges or where overburden surfaces are partially vegetated, photogrammetric surveys capture the top of the vegetation rather than the ground surface. LiDAR can penetrate vegetation to some degree, but the penetration is inconsistent depending on vegetation density and type.

This matters for volume calculations. If your survey captures vegetation tops rather than the true ground surface on a stockpile, your volume estimate will be inflated.

Ground Control Point Coverage

GCPs remain the primary mechanism for achieving reliable absolute accuracy. The accuracy of a drone survey is fundamentally limited by the accuracy and distribution of its ground control. Mines that skimp on GCP maintenance — using old coordinates, insufficient numbers of points, or poorly distributed networks — will get poor results regardless of how good their drone is.

The International Society for Mine Surveying has published guidelines recommending minimum GCP densities for various survey applications in mining. Many operations fall below these recommendations.

Integrating Drones With Traditional Survey Methods

The best mine survey programs don’t choose between drones and traditional methods. They use each where it’s most effective.

Traditional GNSS and total station surveys remain superior for precise control point establishment, boundary surveys with legal significance, and monitoring of specific movement points on pit walls where millimetre-level accuracy matters.

Drones are superior for broad-area coverage, stockpile volumes, progress monitoring, and visual documentation. They’re also invaluable for surveying areas that are unsafe or impractical to access on foot.

The practical approach is to maintain a traditional survey control network, validated regularly, and use it as the foundation for drone surveys. The control network provides the accuracy framework; the drone provides the data density.

Some operations are now integrating AI-powered analysis with their drone survey data to automate change detection between sequential surveys. This means that instead of manually comparing two survey datasets to identify where material has been moved, the system automatically flags areas of significant change and calculates volumes moved. It’s particularly useful for reconciling planned versus actual mining progress.

What’s Coming Next

Several developments are worth watching.

Real-time kinematic (RTK) and post-processed kinematic (PPK) positioning integrated directly into drone platforms is reducing the dependence on GCPs. These systems provide centimetre-level positioning for each image without needing ground markers. They don’t eliminate the need for check points, but they can dramatically reduce the number of GCPs required.

Multi-sensor payloads combining camera, LiDAR, and multispectral sensors on a single platform are becoming lighter and more affordable. This means a single flight can generate photogrammetric imagery for visual documentation, LiDAR for accurate terrain modelling, and multispectral data for geological analysis.

Beyond-visual-line-of-sight (BVLOS) operations, enabled by regulatory changes and improved communication systems, will allow drones to survey larger areas without requiring a pilot within visual range. For large open-pit operations covering several square kilometres, this extends the practical coverage area of a single drone significantly.

The bottom line is that drones are now an essential part of any modern mine survey program. But they work best when operators understand their limitations and use them as one tool in a comprehensive survey approach rather than a replacement for traditional methods.