Drone-Based Subsurface Void, Sinkhole, and Hazard Detection

It can contribute to it. Surface-deformation methods (InSAR, GNSS, prisms) show whether the facility has moved; they do not directly show internal moisture conditions, seepage pathways, or pore-pressure changes. Airborne GPR detects dielectric contrasts associated with moisture variation, internal layering and possible preferential seepage paths inside the embankment and across the beach, adding a non-invasive subsurface screening layer. Reliable results need a baseline survey, repeat acquisition for change detection, and validation against piezometer, borehole or other site-instrumentation data. The operator flies from the crest or solid ground, reducing the need to walk on unsafe tailings surfaces. GPR supplements a GISTM-aligned monitoring program; it does not replace deformation or piezometric monitoring.

Within limits, yes, and when site conditions are favorable. GPR can image shallow voids and disturbed ground beneath apparently stable ground, before mine subsidence becomes visible at the surface, provided the void lies within the antenna's depth range and the overlying ground is dry and low-conductivity enough to transmit the signal. In wet clay, saline soils, or conductive tailings the radar signal may be strongly attenuated and the void may be invisible to GPR, which is when methods such as microgravity, electrical resistivity tomography, seismic investigation, or drilling may be more appropriate. So GPR can be a useful early-screening method in favorable ground, but it is limited in conductive ground and the wrong tool in conductive ground; pre-survey site characterization helps determine whether GPR is suitable or should be combined with other methods.

Penetration depends on the antenna frequency and on the electrical conductivity of the ground. In favorable ground (dry sand, gravel, competent rock with low conductivity), an LF dipole configuration can reach up to 12 meters in soil with relative dielectric permittivity equal to 5 and specific attenuation not exceeding about 5 dB/m. A 500 MHz shielded antenna reaches up to 4 meters in average normal soil at a 100 ns time range. A Zond Aero 600 NG reaches up to 4 meters from the ground in average normal soil, and up to 2 meters with an antenna elevation of 0.6 m.

In conductive ground (wet clay, saline soils, fine-grained mine tailings, mineralized groundwater), real-world penetration is significantly reduced and may fall to only a few dozens of centimeters or even zero, depending on frequency and site conditions. Pre-survey site characterization is required to estimate realistic penetration on a specific site.

GPR is not appropriate when one or more of the following applies.

• The target depth exceeds the practical penetration of the chosen antenna in the local ground.
• The ground is strongly conductive (wet clay, saline soils, conductive tailings, ore-bearing groundwater).
• The target has insufficient dielectric contrast with the surrounding material (for example, a water-filled cavity in saturated soil with similar permittivity).
• The required resolution is finer than centimeter-scale resolution near the surface, or the target footprint is smaller than the antenna footprint at depth.
• The site cannot be flown safely due to obstacles, vegetation, or restricted airspace.

In these cases, ground-coupled GPR, electrical resistivity tomography, seismic methods, microgravity, or direct investigation by drilling may be more appropriate. SPH is open about these limits during pre-sales discussions.

A consistent antenna height above the ground is essential for reliable GPR signal coupling and for reliable depth interpretation.

Variation in elevation changes the recorded two-way travel time to the surface and can distort interpretation if not corrected. In the UAV-based test at Baloži, the GPR was mounted on the rails of a DJI Matrice 350 RTK UAV with a laser-guided True-Terrain Following system to keep the GPR antenna at a steady altitude of 0.5 meters above the ground.

Residual variations are corrected in processing through static correction.

Yes. The drone operator works from solid ground, typically at the embankment crest, with no need to walk on the tailings surface. Airborne ground-penetrating radar (GPR) can support moisture-condition monitoring in tailings storage facilities (TSFs), especially for larger or more remote sites.

Reliable monitoring of internal moisture requires baseline acquisition, repeated surveys for change detection, and validation against borehole or instrumentation data.

GPR and ERT are highly complementary rather than competing methods. GPR is much faster to deploy and delivers high spatial resolution at shallow depths in favorable ground. ERT can also produce very high-resolution subsurface images, with resolution that can be tuned by electrode spacing for the target depth and it performs very well in conductive ground, for example, clay, where GPR signal penetration is limited.

The trade-off is in field effort and interpretation. ERT acquisition is significantly slower than GPR: a single line can take hours rather than minutes, more complex multi-line or 3D surveys typically require multiple field operators, and data processing must be done by an experienced geophysicist who can manage inversion, mesh design, and quality control. In practice, GPR and ERT are routinely combined on the same site: GPR delivers fast, high-resolution shallow imaging across larger areas, and ERT fills in the depth and conductive-ground gaps where GPR underperforms.

Point targets such as buried pipes or small cavities appear as diffraction hyperbolas in radargrams; layered structures appear as continuous horizontal reflectors. Depth conversion requires a site-specific electromagnetic wave velocity, estimated through hyperbola fitting or a known reference target. Interpretation is validated against existing ground truth (boreholes, test pits, known infrastructure).

Where ground truth is absent, GPR results are treated as hypotheses for targeted investigation or follow-up.

For mineralized targets, GPR has significant limitations. Many mineralized environments contain conductive lithologies, alteration, or saline groundwater, which strongly attenuate the GPR signal. Standard exploration geophysics relies more on magnetics, electromagnetics, gravity, and radiometrics.

GPR has a role in shallow regolith and overburdened characterization but is not the primary exploration tool.

Drone operations follow the aviation rules of the country of operation. In the European Union, this includes the EASA framework and Specific-category authorizations for larger or higher-risk operations. In the United States, FAA Part 107 applies, with waivers for beyond-visual-line-of-sight where needed. 

For the GPR method itself, ASTM D6432 in the US provides the standard reference for ground-coupled GPR in subsurface investigation. No equivalent ASTM standard currently exists for airborne GPR specifically, and airborne practice draws on ground-coupled standards adapted to UAV acquisition. 

Mine safety regulations are jurisdictional and operator-specific. Where ICMM principles apply, GISTM provides additional expectations for tailings facility management and monitoring.