Drone geophysics and hyperspectral imaging for mineral exploration

Early-stage mineral exploration has always been a balance between covering maximum ground and managing cost. Historically, that meant helicopter-borne geophysical surveys at $5,000–$15,000 per line-kilometre, followed by expensive ground-truthing crews working through terrain that was sometimes genuinely inaccessible. UAV-mounted geophysical and spectral sensors are disrupting this economics in a fundamental way — offering centimetre-resolution data at a fraction of the cost, with the ability to fly ground clearances that helicopter systems cannot safely achieve.

This article covers the three primary UAV geophysical methods now used in mineral exploration — magnetometry, time-domain electromagnetics, and hyperspectral imaging — with practical guidance on what each detects, which sensor systems are proven, and how to interpret results for the most common commodity targets.

The economics shift in early-stage exploration

A conventional airborne magnetic survey over a 10km × 10km target area might run $180,000–$300,000 for a helicopter-borne system with 50m line spacing. A drone magnetometry survey of the same area with 10m line spacing — providing 5× the spatial resolution — can be completed for $15,000–$35,000 depending on terrain and mobilisation cost. The resolution advantage is decisive: 10m line spacing captures structural details (faults, dyke contacts, lithological boundaries) that are invisible at 50m.

For hyperspectral surveys, the comparison is even more dramatic. Satellite hyperspectral data (e.g., ASTER, Sentinel-2 with limited spectral bands) comes at a coarse 10–30m pixel resolution. Drone hyperspectral over a prospect can deliver 5–20cm pixels with full VNIR-SWIR coverage — the difference between mapping an alteration zone and mapping individual mineralised outcrop faces.

UAV magnetometry — mapping the structure beneath

Magnetic surveys measure variations in the Earth’s total magnetic field caused by differences in the magnetic susceptibility of subsurface rocks. Magnetite-bearing rocks (many mafic volcanics, skarns, iron oxide copper-gold systems) produce positive magnetic anomalies. Felsic rocks and most alteration zones (where magnetite has been destroyed by hydrothermal fluids) produce relative magnetic lows. Faults and dykes produce linear anomalies that trace geological structure.

How drone magnetic surveys work

The magnetometer is suspended below the drone on a rigid or semi-rigid mount, typically 1.5–3m below the airframe, to separate the sensor from the magnetic noise produced by the drone’s motors and electronics. Surveys are flown at 30–80m above ground level (AGL) along parallel flight lines spaced 5–25m apart depending on the target depth and resolution required. GPS position, barometric altitude, and magnetic field readings are logged simultaneously at 10–100 Hz.

Total-field data is corrected for diurnal variation (using a stationary base station magnetometer logging throughout the flight), levelled between flight lines, and then processed with standard magnetics software to generate total-field intensity maps, residual anomaly maps, and derivative products (tilt-angle, analytic signal, vertical derivative) that enhance shallow structure.

Key sensor systems

Caesium vapour sensors are strongly preferred over fluxgate magnetometers for exploration-grade surveys. Fluxgate systems are adequate for detecting large, shallow iron ore bodies but lack the sensitivity to resolve subtle alteration-related magnetic depletion zones critical for gold and copper targeting.

What magnetic surveys find

• Magnetite-destructive alteration halos around gold deposits (potassic, propylitic, phyllic zones)
• Iron oxide copper-gold (IOCG) systems — typically strong positive anomalies
• Nickel sulphide targets hosted in ultramafic bodies (distinct magnetic highs)
• Skarn deposits (contact metamorphic magnetite)
• Faults, shear zones, and structural corridors that control mineralisation
• Kimberlite pipes (often subtle negative anomalies in cratonic settings)
• Sub-basalt mapping where sediment-hosted targets lie beneath volcanic cover

UAV time-domain electromagnetics (TEM)

Electromagnetic methods measure how the ground responds to a pulsed electrical current transmitted through a loop. Conductive bodies (massive sulphides, graphite, saline groundwater) produce secondary EM fields as eddy currents decay after the primary pulse — and these secondary fields are detected by receiver coils. TEM is the primary method for discovering buried massive sulphide deposits.

Traditional ground TEM requires crews to lay out 100–400m transmitter loops manually — days of work in accessible terrain, and effectively impossible in rugged or swampy ground. Drone-towed TEM systems suspend a lightweight transmitter-receiver coil assembly beneath the UAV on a 30–50m tether, allowing the survey to fly over obstacles and access terrain that ground crews cannot.

Current commercial drone TEM systems (such as the SkyTEM and GEOTEM-UAV platforms) are capable of detecting conductors to depths of 100–300m depending on the conductor’s size and conductivity, with anomaly resolution adequate to rank drill targets. The method is particularly effective for:

• Volcanogenic massive sulphide (VMS) targets in greenstone belts
• Sediment-hosted massive sulphide (SHMS) systems
• Nickel sulphide targets associated with ultramafic intrusions
• Conductive graphite horizons (which are drill-hazard false positives, but important to map)
• Submarine massive sulphide analogues in shallow marine exploration

Hyperspectral imaging — reading mineralogy from above

Every mineral has a characteristic spectral reflectance signature — a pattern of wavelengths it absorbs and reflects that acts as a fingerprint. Hyperspectral sensors collect reflected light across dozens to hundreds of narrow, contiguous wavelength bands (vs. the 3–5 broad bands in standard multispectral sensors), generating a full reflectance spectrum for every pixel in the image. This spectrum can be compared against mineral reference libraries to identify specific mineral assemblages at the surface.

For mineral exploration, two wavelength regions are most informative:

VNIR (400–1000nm)

The visible-to-near-infrared range detects iron oxide and iron hydroxide minerals. Haematite, goethite, jarosite, and ferrihydrite all show diagnostic absorption features in this region. Iron oxide gossans — weathered expressions of sulphide bodies — are primary targets. The VNIR range also distinguishes vegetation cover, soil, and bare rock for mapping.

SWIR (1000–2500nm)

The shortwave infrared range is where clay minerals, micas, carbonates, and sulphates express their diagnostic absorption features. This is the core exploration wavelength range for mapping hydrothermal alteration assemblages. Key absorption features:

Sensor systems for exploration-grade hyperspectral

• Headwall Photonics Nano-Hyperspec: 400–1000nm VNIR, push-broom, 270 bands. Widely available on drone platforms; excellent for iron oxide mapping.
• Headwall Photonics SWIR: 900–2500nm, 260 bands. Companion sensor for the Nano-Hyperspec; provides full alteration mapping capability when flown together.
• Resonon Pika XC2: 400–1000nm, 240 bands. Compact, ruggedised for field deployment.
• Resonon Pika IR: 900–1700nm, 164 bands. Partial SWIR coverage.
• SPECIM AFX10: 400–1000nm VNIR. 224 bands, designed for UAV integration.
• SPECIM AFX17: 970–1770nm. Extends coverage into SWIR-I range; used when full SWIR-II (to 2500nm) is not required.

Commodity-specific applications

Gold — epithermal and orogenic deposits

High-sulphidation epithermal gold systems are among the most hyperspectral-amenable targets. The advanced argillic alteration assemblage (alunite + kaolinite + dickite + pyrophyllite) has highly distinctive SWIR signatures and often extends hundreds of metres from the high-grade core, making it an excellent surface mapping target. Low-sulphidation systems show adularia + silica assemblages without strong clay signatures, making them more magnetics-dependent for targeting.

Orogenic gold deposits (greenstone-hosted, structurally controlled) are best targeted with UAV magnetics that trace the fault corridors and shear zones. Alteration (carbonate ± sericite ± chlorite) is detectable in SWIR but is often overprinted by weathering in tropical terrains.

Copper porphyry

Porphyry copper systems develop concentric alteration zones that are almost perfectly mapped by hyperspectral imaging. From core to margin: potassic (biotite + K-feldspar + magnetite) → phyllic (quartz + sericite/illite + pyrite) → propylitic (chlorite + epidote + calcite). UAV hyperspectral can delineate these zones at sub-metre resolution over an exposed porphyry system. The illite Al-OH wavelength position (2195 vs. 2210nm) has been shown to correlate with depth to the potassic core, providing a vector toward the highest-grade mineralisation.

REE — carbonatites and alkaline complexes

Rare earth element deposits hosted in carbonatites and nepheline syenites are challenging exploration targets because they often lack clear geophysical expression. UAV magnetics can map the circular to elliptical structural anomalies associated with alkaline intrusions. Hyperspectral imaging detects carbonate mineralogy (calcite, dolomite, siderite) and certain REE-bearing phases (bastnäsite has weak but detectable absorptions) to delineate prospective carbonatite bodies from surrounding lithologies.

Lithium — pegmatites and brine systems

Spodumene-bearing lithium pegmatites are detectable by UAV hyperspectral through the silica, mica (lepidolite at 2350nm), and clay signatures of the altered margins. Pegmatite body geometry (elongated, discordant to regional foliation) is also visible in UAV magnetic and structural mapping. For lithium brine targets, drone magnetics helps map basin structure and fault conduits but SWIR hyperspectral has limited direct utility in dry lake environments.

Processing workflow

Raw hyperspectral data requires significant processing before mineral identification is possible:

• Radiometric calibration: raw DN values converted to at-sensor radiance using sensor calibration files
• Atmospheric correction: QUAC (Quick Atmospheric Correction, empirical, no ancillary data required) or FLAASH (uses standard atmospheric model + sensor altitude + sun angle). FLAASH is preferred for quantitative mineral mapping.
• Geometric correction: GPS/IMU data + ground control points used to orthorectify the image to a consistent ground coordinate system
• Reflectance spectra extraction: per-pixel spectra extracted from the corrected image cube
• Mineral identification: Spectral Angle Mapper (SAM) or Matched Filter algorithms compare pixel spectra against reference mineral library spectra (USGS Spectral Library, ASTER Spectral Library)
• Alteration zone mapping: classified mineral maps colour-coded by assemblage and visualised against topography

Standard software for hyperspectral processing includes ENVI (Harris Geospatial), SpecTIR, and open-source tools including HyPy and Python spectral libraries. Pilots who deliver raw radiance cubes without atmospheric correction are providing data that cannot be used for quantitative mineral identification — always specify fully processed, atmospherically corrected reflectance cubes as your deliverable.

Commissioning a drone geophysics survey

When creating a geophysics mission on Vantagr, the mission brief should specify:

• Survey type: magnetics, TEM, or hyperspectral (specify VNIR, SWIR, or both)
• Target area: polygon with area in km². Remote areas require mobilisation planning.
• Line spacing and AGL: tighter spacing and lower altitude = higher resolution but more flight time and cost
• Magnetic datum: base station location for diurnal correction (magnetics only)
• Deliverable format: for magnetics, XYZ ASCII + gridded GeoTIFF; for hyperspectral, orthorectified reflectance cube (.hdr/.img or ENVI format) + classified mineral map GeoTIFF
• Processing level: raw data only, or full processed deliverable including mineral classification
• Coordinate system and datum: UTM zone + WGS84 or local datum