Hyperspectral remote sensing for mineral exploration measures reflected sunlight in hundreds of narrow, contiguous bands, so each pixel carries a near-continuous spectrum. Many alteration minerals have diagnostic absorption features in that spectrum: white mica and kaolinite near 2,200 nm, iron oxides in the visible and near-infrared, and rare-earth elements such as neodymium at sharp features between about 580 and 870 nm. Multispectral sensors see a few broad bands and can only hint at these minerals. Hyperspectral sensors can often tell them apart.

Try it: Hyperspectral data volume and swath planner →

That is the promise. The limits are just as important: pixels of 30 to 60 m, vegetation and cover that hide the rocks, and the simple fact that satellites read the surface, not the orebody. We set out what each sensor offers, what the published studies show for gold and rare earths, what happened to ASTER, and how to plan a survey sensibly. Figures come from mission pages and peer-reviewed papers, and we say where evidence is thin.

“EMIT records 285 spectral bands from the International Space Station at 60 m resolution.”

How hyperspectral remote sensing for mineral exploration works

Every mineral absorbs light at particular wavelengths because of the bonds and ions in its structure. Hydroxyl bonded to aluminium (Al-OH) absorbs near 2,200 nm; hydroxyl bonded to magnesium or iron absorbs further out, around 2,250 to 2,350 nm; ferric iron absorbs in the visible and near-infrared. A spectrometer that samples every few nanometres can resolve those features and match them against reference spectra, such as those in the USGS Spectral Library.

A 2023 study in Scientific Reports, “Towards better delineation of hydrothermal alterations via multi-sensor remote sensing and airborne geophysical data”, summarises the features explorers rely on. A sharp Al-OH trough near 2,200 nm with a minor feature near 2,350 nm points to muscovite, a guide to phyllic (sericite) alteration. Kaolinite, typical of argillic alteration, shows a major feature at 2,200 nm with a minor one at 2,170 nm. Epidote in propylitic zones has a shallow feature at 2,250 nm, and limonite absorbs strongly at 480 nm with a secondary trough at 930 nm.

Why band count matters

The two kaolinite features at 2,170 and 2,200 nm sit only 30 nm apart. A multispectral band tens of nanometres wide averages across both features and cannot separate kaolinite from white mica or smectite with confidence. A hyperspectral sensor sampling every 10 nm or so in the shortwave infrared can. That difference is the whole case for hyperspectral imaging in mining: better mineral discrimination means alteration zoning can be mapped, not just flagged.

The satellites: EnMAP, PRISMA, EMIT and the ASTER archive

Several spaceborne sensors now make hyperspectral remote sensing for mineral exploration practical over whole licences, rather than only over small airborne survey blocks. Each involves trade-offs between pixel size, swath and band count.

EnMAP (Germany)

The EnMAP mission covers 420 to 1,000 nm in the VNIR and 900 to 2,450 nm in the SWIR, with 30 m by 30 m pixels over a 30 km swath. Its nadir revisit is 27 days, reduced to 4 days with off-nadir pointing. The instrument specification gives spectral sampling of 6.5 nm in the VNIR and 10 nm in the SWIR, with a signal-to-noise ratio above 150 at 2,200 nm, right where the clay and mica features sit.

PRISMA (Italy)

PRISMA, built and operated by the Italian Space Agency and launched on 22 March 2019, carries a hyperspectral camera with 239 channels (66 from 400 to 1,010 nm and 173 from 920 to 2,505 nm) plus a panchromatic channel. Pixels are 30 m over a 30 km swath, with spectral resolution better than 12 nm.

EMIT (NASA, on the Space Station)

EMIT launched to the International Space Station on 14 July 2022. It has 285 spectral bands, 60 m pixels and a 75 km swath. It was built to map the mineralogy of dust source regions in the world’s arid zones, which also makes its data useful for geology in deserts, although the 60 m pixel is coarse for small targets.

ASTER: the multispectral workhorse, now mostly an archive

ASTER on NASA’s Terra satellite is multispectral, not hyperspectral, but it shaped satellite alteration mapping for two decades. NASA’s instrument page lists 14 bands: VNIR at 15 m, six SWIR bands at 30 m and five thermal bands at 90 m. Its SWIR detectors failed in 2008; the LP DAAC advisory says SWIR data acquired since April 2008 are not usable. NASA’s ASTER transition notice adds that the thermal subsystem was permanently turned off on 16 January 2026 because of Terra power limits, leaving only VNIR collecting. The archive stays available.

What comes next

Europe’s Copernicus programme is building CHIME, a two-satellite hyperspectral mission. The CHIME page on SentiWiki lists a 30 m ground resolution over a 130 km swath, raw materials among its policy uses, and working launch dates of 2028 for CHIME-A and 2030 for CHIME-B. A wide swath at 30 m would make regional hyperspectral coverage far more routine than today’s 30 km strips.

Hyperspectral gold detection: mapping the alteration halo

Searches for hyperspectral gold detection accuracy are common, and the honest answer starts with what is being detected. Gold grains are far too small and sparse to produce a spectral signature at 30 m. What hyperspectral remote sensing for mineral exploration maps is the alteration that hydrothermal fluids leave around many gold deposits: sericite and other white micas, kaolinite and other clays, silica, carbonates, and iron oxides where sulphides have weathered.

In epithermal and porphyry-related gold systems, alteration is often zoned: argillic and advanced argillic assemblages near the top or core, phyllic sericite zones, and propylitic chlorite-epidote halos further out. Mapping that zoning from orbit tells geologists where a system’s centre might sit. In orogenic gold belts the alteration is usually narrower and follows shear zones, so structure mapping and multispectral data carry more of the load.

What “accuracy” means here

There is no single accuracy figure for satellite gold detection, because the satellite does not detect gold. Published studies report how well mapped alteration matches field or laboratory mineralogy, and that varies with exposure, vegetation, pixel size and the minerals involved. The useful question is narrower: does the alteration map match what geologists find on the ground at a sample of sites? Any provider should be able to show that validation.

• ✔ Strong use: mapping clay, mica and iron-oxide zoning in well-exposed, arid or semi-arid terrain.
• ✔ Strong use: screening a large licence to decide where mapping and sampling crews go first.
• Moderate use: partly vegetated ground, where spectra mix with plant signatures.
• ⚠ Weak use: dense forest, deep transported cover or thick laterite, where the rocks are not visible.

For how satellites fit a gold programme more broadly, see our guide to satellite gold exploration, which covers multispectral screening, structure and field follow-up.

Rare earths: where hyperspectral imaging in mining has a direct signal

Rare-earth elements are the unusual case for hyperspectral remote sensing for mineral exploration. Trivalent neodymium produces sharp absorption features in the visible and near-infrared, roughly at 580, 740, 800 and 870 nm, and their positions do not depend much on the host mineral. That gives hyperspectral remote sensing for mineral exploration a more direct signal for REE deposits than for most metals.

A 2024 study by Saeid Asadzadeh, Nicole Koellner and Sabine Chabrillat, “Detecting rare earth elements using EnMAP hyperspectral satellite data: a case study from Mountain Pass, California”, used the ~740 and ~800 nm features to map neodymium at EnMAP’s 30 m pixel size over the bastnaesite-bearing carbonatite. The authors cite laboratory detection limits for Nd ranging from about 1,000 ppm down to below 200 ppm, and state plainly that the smallest REE-bearing target detectable at 30 m is not yet known.

• 📊 Carbonatites and alkaline intrusions are the classic hard-rock REE targets where outcrop allows spectral mapping.
• 📊 Neodymium is a proxy: it is the element with the clearest features, and its presence suggests other light rare earths in the same minerals.
• Validation matters: the Mountain Pass study compared its maps with laboratory spectra of ore samples and high-resolution imagery.
• ⚠ Clay-hosted REE deposits are a different problem: the rare earths are adsorbed on clays, without the same sharp spectral features.

Planning a hyperspectral survey of your licence

Good hyperspectral remote sensing for mineral exploration starts with logistics. Hyperspectral data is heavy. A single 30 m scene with over 200 bands holds many times the data of a multispectral scene of the same area, and narrow swaths mean a large licence may need several strips. Before ordering or requesting acquisitions, it helps to know roughly how many strips and how much data are involved.

Practical steps

1. Check the archive first. EnMAP, PRISMA and EMIT all hold archived scenes; your licence may already be covered.
2. Pick the season. Dry-season, low-vegetation scenes with the sun reasonably high give the cleanest mineral spectra.
3. Correct the atmosphere. Work in surface reflectance, not top-of-atmosphere radiance, or water and carbon dioxide bands will distort the SWIR.
4. Mask what is not rock. Remove vegetation, water, cloud shadow and infrastructure before mapping minerals.
5. Map, then validate. Check a sample of mapped sites in the field or with a handheld spectrometer before ranking targets.

From orbit to core tray: the hyperspectral scale ladder

Satellites are only the top rung. Hyperspectral remote sensing for mineral exploration works best as a ladder, where each scale checks and sharpens the one above it. The same absorption features are measured at every step; what changes is the footprint.

Satellite: tens of metres, whole licences

At 30 to 60 m, a pixel mixes rock, soil, vegetation and shadow. Satellite data is best at finding where alteration minerals cluster across a licence or district, and at showing zoning patterns hundreds of metres to kilometres across. It is the cheapest way to decide where the rest of the ladder should go.

Airborne: metres, selected blocks

Airborne imaging spectrometers flown on aircraft or drones record the same spectral range with pixels of metres or less. They cost more per square kilometre and need permits and flight planning, so they are usually flown over the best satellite targets rather than whole licences. At that resolution individual veins, alteration selvages and small outcrops start to resolve. Our guide to drone geophysics and LiDAR surveys covers the flying side.

Field and laboratory: centimetres, validation

Handheld spectrometers measure a rock or soil surface directly, and give the ground truth that satellite and airborne maps are checked against. Core and chip scanners bring the same measurement into the drilling phase, logging alteration mineralogy downhole. When the satellite map, the field spectra and the core agree, a target becomes much easier to believe.

The practical lesson is to plan the ladder from the start. A satellite alteration map with no field spectra behind it is a hypothesis; the same map checked at a few dozen sites is evidence. Budget a handful of field days for validation in any programme that relies on hyperspectral remote sensing for mineral exploration.

ASTER, multispectral and hyperspectral compared

ASTER is still the reference point many geologists start from, so it is worth being clear about its status. The step chart shows the number of usable ASTER bands since launch.

Remote sensing’s two classic jobs in exploration, as Sabins’ review in Ore Geology Reviews describes, are mapping the structures that localise ore and recognising hydrothermally altered rocks by their spectra. Hyperspectral data improves the second job; multispectral and radar data still do much of the first. Our guide to remote sensing methods for mineral exploration covers both.

Questions to ask before you buy a hyperspectral study

Whether you commission a study or buy processed maps, the same questions separate careful hyperspectral remote sensing for mineral exploration from colourful pictures. Ask them in writing.

1. Which sensor, which dates, and why? Scene dates matter for vegetation, moisture and sun angle.
2. How was the atmosphere corrected? Reflectance products should be named, with the method used.
3. What was masked out? Vegetation, water, cloud, shadow and infrastructure should be removed before mineral mapping.
4. Which reference spectra were used? A named library or field spectra from the site, not an unexplained colour scale.
5. How were results validated? Field checks, handheld spectra, laboratory mineralogy or assays at a stated number of sites.
6. What are the limits? Minimum mappable target size, confidence per mineral class, and areas where the method does not work.

A provider who answers these clearly is also telling you how much weight the maps can carry in a drilling decision or an investor presentation.

How we use hyperspectral remote sensing for mineral exploration

Our satellite-based mineral detection analyses reflected electromagnetic energy, using multispectral or hyperspectral data depending on the project, to flag likely mineralised target zones, alteration halos, faults and fractures. For each project we pick the data source that suits the area and target mineral, acquire the data, run the analysis and deliver the results. We have scanned more than 100,000 hectares for over 20 mineral types across 25+ countries. Detectable materials include gold, silver, copper, lithium and rare earth elements.

• What you send: coordinates, a KML/KMZ file or a polygon, plus country and target mineral.
• Turnaround: 5–20 business days depending on area and mineral complexity.
• 📊 Premium report: high-potential zones, prospectivity heatmaps, estimated location and depth ranges, indicative quantity, geological interpretation and seasonal anomaly validation, as a PDF plus georeferenced GIS files.
• 📊 Premium+ report: adds TargetMax™ Drilling Intelligence with drilling-angle recommendations and interactive 3D subsurface models.
• Economics: exploration timelines from months to days, and up to 80–85% lower early-exploration cost.
• ⚠ Limits: targets are exploration targets, not resources, and must be confirmed by sampling and drilling.

See a sample of satellite-driven 3D mineral prospectivity mapping, request pricing through our mining query form, or draw your boundary on mining.farmonaut.com: Map Your Mining Site. The wider exploration sequence is in our mineral exploration guide.

Frequently asked questions

Reviewed September 2026 against the EnMAP mission page and instrument specification, eoPortal’s PRISMA entry, NASA Earthdata’s EMIT feature and ASTER instrument page, the LP DAAC ASTER SWIR advisory, NASA’s ASTER transition notice, the Copernicus SentiWiki pages for CHIME and Sentinel-2, USGS pages on Landsat 8 and the USGS Spectral Library, the 2023 Scientific Reports multi-sensor alteration study, Asadzadeh, Koellner and Chabrillat (2024) on EnMAP REE mapping, and Sabins (1999).

Mission parameters and launch dates change; check the mission pages before planning acquisitions. Satellite targets and spectral mineral maps are exploration targets, not mineral resources, and must be confirmed by sampling and drilling.