Structural geology and lineament mapping for mineral exploration
Satellite SAR, optical and elevation data expose fault traces, dyke swarms and lithological contacts across terrain that would take field crews years to walk. The method builds structural frameworks that sharpen drill-target selection, with honest limits on depth projection.
Sensors
- ALOS-2 PALSAR-2 (JAXA): L-band SAR at 1.27 GHz; 3 m resolution in spotlight mode, 10 m in stripmap. Long wavelength penetrates dry overburden and thin vegetation, exposing structural grain in radar backscatter texture that shorter wavelengths miss. Repeat cycle 14 days.
- Sentinel-1 A/B (ESA): C-band SAR at 5.405 GHz; 5 x 20 m IW mode. Six-day revisit with two satellites. Useful for lineament extraction via backscatter texture and for stacking multi-temporal images to suppress noise; less penetrating than L-band in vegetated terrain.
- TanDEM-X / SRTM DEM: TanDEM-X global DEM at 12 m posting (0.4 arcsec) with relative vertical accuracy around 2 m. SRTM at 30 m. Both expose topographic expression of faults: fault scarps, offset drainage, pressure ridges. Hillshading from multiple azimuths reveals lineaments invisible in a single illumination direction.
- ASI PRISMA: Hyperspectral-panchromatic sensor, 30 m spatial resolution, 400–2500 nm in 239 bands. Lithological contacts and alteration halos that bound structural corridors are detectable spectrally. Note: mineral mapping from PRISMA is covered in a sibling page; here it serves as a contact-delineation complement to structural mapping.
What radar backscatter actually reads in the rock
A SAR image is not a photograph. The sensor measures how rough a surface is at the wavelength of the radar pulse. Structural geology leaves its signature in that roughness: a fault scarp produces a sharp backscatter boundary; a resistant dyke standing proud of softer country rock creates a bright linear return; foliation in metamorphic belts imposes a repeating texture that aligns with regional strike. These signals are independent of illumination angle, cloud cover and, to a useful degree, surface vegetation.
L-band systems such as ALOS-2 PALSAR-2 are particularly valuable in exploration because the 23 cm wavelength interacts with the soil and shallow regolith rather than just the canopy. In tropical terrains where optical sensors see only tree canopy, L-band can still resolve the structural grain of the underlying rock mass. C-band Sentinel-1 is more susceptible to vegetation scattering but remains effective in arid and semi-arid terrain, which covers a large share of the world's under-explored mineral belts.
Topography as a structural archive
Faults that have been active recently enough to leave topographic expression are visible in any good DEM. The technique is straightforward: generate hillshade rasters from multiple solar azimuths, typically at 45-degree intervals around the compass, then stack or composite them. Lineaments that appear in only one illumination direction are probably artefacts of the DEM or agricultural features; those that persist across azimuths are structural candidates.
The TanDEM-X 12 m DEM resolves scarps of a few metres vertical displacement across most of its global coverage. SRTM at 30 m is coarser but covers regions where TanDEM-X data are not yet commercially released. Drainage deflection, wind gaps and faceted spurs are secondary indicators that a trained analyst reads from the same surface. None of this tells you how deep the fault penetrates or whether it is mineralised. It tells you where to start.
Extracting lineaments: method and its honest limits
Automated lineament extraction applies edge-detection filters (Canny, Sobel, or directional Gabor kernels) to SAR amplitude images and DEM derivatives, then applies length and straightness thresholds to reject noise. Rose diagrams of the surviving lineaments reveal dominant structural trends and can be compared against regional tectonic compilations. Published studies on PALSAR and Sentinel-1 data in gold and copper terrains have demonstrated detection of lineaments down to roughly 500 m in length at 10 m resolution, with shorter features increasingly contaminated by agricultural field boundaries and road networks in cultivated areas.
The interpretive limit is significant and should be stated plainly. A lineament on a satellite image is a hypothesis, not a mapped fault. Some lineaments are lithological contacts, not faults. Others are erosional or depositional features with no structural significance. Ground-truthing, geophysical surveys (magnetics, gravity) and existing geological maps are necessary to discriminate between them. The satellite product narrows the search space; it does not replace the geologist.
Depth projection is the most consequential uncertainty. Ore-forming fluids circulate at depths of hundreds to thousands of metres. A fault visible at surface may be a shallow splay with no connection to deep fluid pathways, or it may be the surface expression of a crustal-scale structure that channelled mineralisation. Satellite data cannot resolve this ambiguity. Integration with aeromagnetic depth-to-basement models and seismic reflection data is required before structural lineaments from imagery can be ranked as drill targets with any confidence.
Multi-sensor fusion: where the value compounds
No single sensor covers the full structural picture. A practical workflow stacks at least three inputs: a SAR amplitude image for backscatter texture, a DEM hillshade composite for topographic expression, and an optical or hyperspectral image for colour and spectral contrast at lithological contacts. Where all three agree on a linear feature, confidence in a structural interpretation rises substantially.
PRISMA's 239-band hyperspectral coverage at 30 m resolution adds a further dimension. Alteration minerals associated with hydrothermal systems, particularly phyllosilicates, carbonates and iron oxides, have diagnostic absorption features in the shortwave infrared that PRISMA resolves. When a structural lineament from SAR and DEM analysis coincides spatially with an alteration anomaly from PRISMA, the combined signal is a materially stronger exploration lead than either dataset alone. The spatial resolution of PRISMA (30 m) means that narrow alteration halos less than roughly 60 m wide may be missed or smeared; this is a real constraint in terrains with tight structural corridors.
Archive depth matters for regional work. Sentinel-1 data are freely available from 2014 onwards through the Copernicus Data Space. ALOS-2 archive access is negotiated through JAXA or commercial resellers. SRTM data from the 2000 shuttle mission remain the most consistent global baseline DEM. For greenfield exploration in remote regions, the ability to build a structural framework before a single geologist boards a plane is the primary economic argument for this approach.
From structural map to drill-target ranking
A structural framework on its own does not rank targets. The ranking step combines lineament density maps, intersection nodes (where two or more fault trends cross, a classic locus for fluid focusing), proximity to known intrusions or favourable host lithologies, and any available geochemical or geophysical layers. The output is a prioritised list of structural corridors with explicit confidence tiers based on how many independent datasets converge.
Satellize runs this multi-layer analysis on open constellations and commercial tasking under client licence. The workflow is the same one the company applies in the Tonga crop-estimation programme at a different scale: ingest, fuse, quantify, deliver a GIS layer the client's geologists can interrogate directly rather than a static PDF. The deliverable for structural geology work is a vector lineament dataset with attribute tables recording which sensors detected each feature, its length, azimuth and intersection count, alongside a ranked target polygon layer and a written interpretation note that states assumptions and uncertainties explicitly.
Typical figures
| Best SAR spatial resolution (spotlight) | 3 m (ALOS-2 PALSAR-2); 5 x 5 m (Sentinel-1 SM mode) |
| Operational SAR resolution (stripmap / IW) | 10 m ALOS-2 stripmap; 5 x 20 m Sentinel-1 IW |
| DEM posting | 12 m TanDEM-X; 30 m SRTM; relative vertical accuracy ~2 m (TanDEM-X) |
| Hyperspectral resolution | 30 m (PRISMA), 239 bands, 400–2500 nm |
| SAR revisit | 6 days (Sentinel-1 two-satellite); 14 days (ALOS-2) |
| Minimum detectable lineament length | ~500 m reliably at 10 m SAR resolution; shorter features require manual review |
| Cloud sensitivity | SAR and DEM: none. Optical/hyperspectral: cloud-free acquisition required |
| Archive depth | Sentinel-1: 2014–present (free). SRTM DEM: 2000 (free). ALOS-2: 2014–present (licensed) |
| Delivery formats | GeoTIFF (raster products), GeoPackage / Shapefile (vector lineaments), CSV attribute tables, PDF interpretation note |
Analytics Satellize can run
| Lineament extraction map | Edge-detection filtering (Canny / Gabor) on SAR amplitude and DEM hillshade composites; automated vectorisation with length and straightness thresholds | Vector GeoPackage of lineaments with azimuth, length and multi-sensor confidence attributes |
| Rose diagram and structural trend analysis | Circular statistics on lineament azimuths; comparison against published regional tectonic compilations | PDF report with rose diagrams, dominant trend summary and tectonic context note |
| Lineament intersection density grid | Kernel density estimation on lineament intersection nodes; rasterised to 50–100 m grid | GeoTIFF density raster and ranked intersection-node point layer |
| Multi-azimuth DEM hillshade composite | Hillshading at 8 azimuths from TanDEM-X or SRTM; RGB composite of three selected azimuths for visual interpretation | GeoTIFF composite and individual azimuth hillshades |
| Structural-alteration coincidence layer | Spatial intersection of SAR/DEM lineaments with PRISMA-derived alteration index rasters (SWIR band ratios for clay, carbonate, iron oxide) | Vector polygon layer of coincident structural-alteration anomalies, ranked by sensor convergence count |
| Drill-target structural ranking | Multi-criteria scoring: lineament density, intersection count, alteration coincidence, proximity to mapped contacts; uncertainty tier assigned per target | Ranked target polygon GIS layer with scoring table and written interpretation note stating assumptions and depth-projection limits |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.