What kinds of sensors watch the Earth from orbit, what can each see, and what is each used for?
The short answer. six sensor families matter commercially. Optical cameras see what the eye sees, at resolutions down to about 30 cm commercially and 10 cm in very low orbit, but only in daylight and without cloud. Synthetic aperture radar (SAR) sees through cloud and at night, measures tiny movements of the ground and finds ships and vehicles, at 25 cm to a few metres. Hyperspectral sensors split light into hundreds of bands to identify materials, crops and gases such as methane. Thermal infrared measures heat. Radio-frequency (RF) sensing geolocates transmitters such as ships' radars and phones. Atmospheric sounders feed weather forecasting. Each family has different economics: optical has the largest market and the steepest price competition; SAR is growing fastest; hyperspectral, thermal and RF are niches with specialised buyers.
The families
| Sensor | What it measures | Resolution (commercial) | Works at night or through cloud? | Main uses | Example operators |
|---|---|---|---|---|---|
| Optical (panchromatic and multispectral) | Reflected sunlight in visible and near-infrared | ~30 cm (best commercial), 3–5 m (monitoring), 10 m (Sentinel-2) | No | Mapping, defence, agriculture, insurance, urban change | Vantor (Maxar), Planet, Airbus, BlackSky, Satellogic, Chang Guang |
| SAR | Radar echoes; phase for movement | ~25 cm to several metres | Yes | Defence, maritime, flood and disaster, ground subsidence, infrastructure | ICEYE, Capella, Umbra, Airbus, iQPS, Synspective |
| Hyperspectral | Hundreds of narrow spectral bands | ~5–30 m | No | Minerals, crop health, methane and pollution | Pixxel, Planet/Carbon Mapper (Tanager), Wyvern, EnMAP (Germany) |
| Thermal infrared | Surface temperature | ~3.5–100 m | Partly at night | Water use, urban heat, industrial activity, fires | Hydrosat, constellr, Satellite Vu, Landsat |
| RF geolocation | Radio emissions | Location accuracy hundreds of metres to km | Yes | Dark vessels, GPS jamming, spectrum monitoring | HawkEye 360, Unseenlabs, Kleos |
| Atmospheric and weather | Radio occultation, sounders, AIS | Kilometres | Yes | Weather forecasting, ship tracking | Spire, Tomorrow.io, PlanetiQ |
How the families complement each other
Optical answers "what is there"; SAR answers "what changed, even at night or under cloud"; hyperspectral and thermal answer "what is it made of" and "how hot is it"; RF answers "who is transmitting". Most serious users combine them. A navy looking for a ship that has switched off its transponder uses RF to find a radar emitter, SAR to see the hull, and optical to identify it. An insurer assessing flood damage uses SAR during the storm and optical after it.
Why resolution is not everything
Higher resolution costs more per square kilometre and covers less area per pass. For many applications, such as crop health, forest cover and large-scale change, 3–10 m resolution with daily revisit is more useful than 30 cm once a month. Chapter 5 sets out the trade-off between resolution, revisit and latency.
Market shares by sensor (Setcoin judgement)
Exhibit 3.1
Optical earns most commercial EO revenue; SAR is the fastest-growing share
Share of commercial EO data revenue, 2025 (indicative) · Market shares by sensor (Setcoin judgement)
Fastest: Optical, ~55–65% a year, doubling in about 1.5 years. Slowest: Hyperspectral and thermal, ~3–5%, doubling in about 18 years
Low endHigh end of the range“+”: at least this% a year · click a segment to follow it
View the data table (5 rows)
| Sensor | Share of commercial EO data revenue, 2025 (indicative) | Trend |
|---|---|---|
| Optical | ~55–65% | Growing slowly; price pressure |
| SAR | ~20–25% | Fastest growing |
| RF and other signals | ~5–10% | Growing with maritime and defence |
| Hyperspectral and thermal | ~3–5% | Early; growing from a small base |
| Weather and atmospheric data | ~5% | Government-led buying |
These are Setcoin judgements based on company revenues and contract announcements; no public source splits the market this way.
What would change this view
A cheap, reliable very-high-resolution SAR at scale would erode optical's share in defence. Mass-market hyperspectral (for example for agriculture or methane regulation) would turn a niche into a large segment.