Satellite Image Resolution: How to Choose the Right Spatial Resolution

Buy the wrong satellite image resolution and one of two things happens: you overpay for detail the project never needed, or you buy cheap wide-area data and still can't answer the question you started with. Matching pixel size to the actual task, not the highest number available, is what makes satellite monitoring worth the cost.

What spatial resolution actually means

Spatial resolution is one pixel's footprint on the ground. A 30 cm resolution satellite image means each pixel represents a 30 by 30 centimeter square; a 2 m resolution image means each pixel covers 2 by 2 meters. Finer resolution (a smaller number) shows more detail per pixel, and costs more per square kilometer to collect. Satellite image resolution is not the only variable that matters, though it's the one most buyers ask about first. Three others shape what a satellite dataset can actually tell you:

• Spectral resolution: how many separate wavelength bands the sensor records, from a single panchromatic band to hundreds of hyperspectral channels.
• Temporal resolution: how often the satellite revisits the same location, from daily commercial tasking to a multi-week cycle for wide-swath sensors.
• Radiometric resolution: how many shades of brightness a pixel can record, which determines how well the sensor separates subtle differences, such as slightly different water turbidity or vegetation stress.

The rest of this guide works through all four, starting with the one that decides most purchasing decisions: spatial resolution.

What is sub-meter resolution satellite imagery?

Sub-meter resolution is any satellite image where one pixel covers less than 1 meter of ground, roughly a person's arm span. It's the umbrella term for the three finest tiers on this page: super high (25–30 cm), very high (31–50 cm), and high (51–80 cm) resolution. Anything coarser than 1 m per pixel, including the 2 m and wider-swath tiers further down, falls into meter-level resolution instead.

That 1 m line is not just a marketing label, it's roughly the point where a satellite image stops resolving individual objects and starts only showing their aggregate shape. The table below shows exactly where specific features drop out as pixel size grows from 0.30 m to 8 m.

The practical takeaway: a wide field or forest boundary doesn't need sub-meter resolution to track, but a construction trench, a parked vehicle, or a lane marking disappears entirely once pixel size passes roughly 1 m. That's the line separating the three sub-meter tiers below from the moderate and wide-swath tier that follows them.

The six satellite resolution tiers, from 15 cm to 50 m

Every commercial and open-access satellite on the market falls into one of six practical resolution tiers. Each trades spatial detail for coverage and cost in a different way.

Satellite image resolution comparison: all six tiers side by side

Key decision factors for choosing satellite image resolution

Picking the resolution of satellite images for a monitoring program comes down to four practical questions, not the sharpest number on the price list.

• Target feature size: if the smallest thing you need to see is a vehicle, foundation, or pipeline valve, choose 0.3–0.5 m. For crop fields or forest blocks, 2–10 m is sufficient. For mineral alteration zones or regional water basins, 30 m hyperspectral or wide-swath data works best.
• Weather and cloud cover: in tropical or flood-prone regions where cloud regularly blocks optical sensors, choose SAR radar imagery (1–3 m) instead of waiting for a clear-sky optical pass.
• Archive vs. new tasking: archived data (90+ days old) is the most affordable baseline for historical change detection. New satellite tasking is required when current ground conditions or a specific capture date matter.
• Minimum order rules: very high and super high resolution optical tasking generally requires a minimum purchase area of 25 km² for archive data and 100 km² for new collections.

Archive and new tasking are not an either-or choice. Most monitoring programs combine both: pull archive imagery first to establish a historical baseline and confirm the resolution actually shows what you need, then layer in new tasking for the current date once the project is live. Starting with cheaper archive data before committing to a tasking order also protects the budget if the required satellite images resolution turns out to be finer, or coarser, than first assumed.

Spectral resolution: what each pixel actually sees

Spectral resolution is how many separate wavelength bands a sensor records, and it works independently of spatial resolution, a 30 m pixel can still carry hundreds of spectral bands.

• Panchromatic: a single band across the visible spectrum, producing a sharp grayscale image with no color information.
• RGB: three bands, red, green, and blue, giving natural color but no ability to separate materials that look the same to the human eye.
• Multispectral: typically 4 to 10 bands, including near-infrared, enough to separate vegetation, water, and built surfaces reliably.
• Hyperspectral: hundreds of narrow, contiguous bands (GF-5B captures 330, spanning 400–2500 nm), fine enough to identify specific minerals, crop stress, and material composition rather than just broad land cover classes.

As a rule, spectral and spatial resolution trade off against each other: packing in more bands generally means a coarser pixel, which is exactly why hyperspectral sensors sit in the 30 m tier rather than the 30 cm tier.

Temporal resolution: how often you get a new image

Temporal resolution is the revisit time, how often a satellite images the same location again. It depends on orbit, swath width, and whether the satellite can be tasked to point off-nadir. Most Earth observation satellites fly in a polar orbit, and revisit times across the industry span anywhere from 1 to 16 days for a single satellite. Super high resolution satellites typically revisit daily; wide-swath sensors like GF-6 or Sentinel-2 revisit every few days over the same tile; hyperspectral and geostationary sensors trade daily coverage for either global reach or continuous same-spot monitoring (GF-4 re-images every 20 seconds from geostationary orbit).

How much temporal resolution a project needs depends entirely on how fast the thing being monitored actually changes. Tracking a decade of urban sprawl tolerates a months-long revisit gap; tracking an active flood or wildfire does not, which is why disaster response leans on daily-tasked or SAR constellations rather than a slower wide-swath satellite.

Spatial and temporal resolution generally trade off too. A narrow, high-resolution sensor covers less ground per pass, so it takes longer to revisit the same tile than a coarser, wide-swath sensor does. Large constellations of smaller satellites, rather than a single satellite, are how commercial operators get around this, using multiple spacecraft to deliver both fine detail and frequent revisit.

Radiometric resolution: the detail inside each pixel

Radiometric resolution is the number of brightness levels a pixel can record, expressed in bits. A 1-bit image only separates black from white; each added bit doubles the number of shades available, up to the 65,536 levels a 16-bit sensor can record. That extra headroom is what separates subtle differences an 8-bit image would flatten into the same shade, like slightly different water turbidity, thin cloud edges, or early-stage vegetation stress. Most current commercial optical satellites capture at 11-bit or higher, then deliver imagery in 8-bit or 16-bit formats depending on the analysis workflow.

Recommended satellite resolution by industry

Construction and real estate

A construction site or property portfolio needs enough detail to catch a specific problem, not just confirm a building exists. At 25 to 50 cm, teams track foundation pours, material stockpiles, and structural progress from a desktop instead of a site visit, and can flag if a contractor's footprint has crept past the boundary line.

Insurance

Claims teams move faster with resolution fine enough to tell partial damage from total loss without sending an adjuster into a hazard zone. Pre- and post-event imagery at 25 to 50 cm resolves roof damage, debris fields, and flood lines building by building, turning a multi-week field survey into a desk review.

Mining and natural resources

Open-pit operations combine two resolution tiers rather than one. Sub-meter imagery tracks haul roads, stockpile growth, and equipment movement day to day, while 30 m hyperspectral data maps the alteration zones and mineral signatures that guide where the next pit expansion or drill program should go.

Energy, utilities, and pipelines

A pipeline or transmission corridor is long, exposed, and rarely staffed, so the resolution has to catch encroachment and vegetation growth along the whole route, not just at inspection points. Imagery at 50 cm to 1.5 m, refreshed on a set schedule, turns a corridor walk that once took weeks into a single scheduled pass.

Agriculture and forestry

Crop and timber monitoring split into two jobs at two different scales. Field-level decisions, spotting a stressed block or an early pest outbreak, need 50 cm to 2 m and a Red Edge or near-infrared band. Regional yield forecasting and forest-cover tracking across a whole province run fine on 10 m open-access data.

Environmental monitoring

One monitoring team can be responsible for a coastline, a wetland, and a forest reserve at once, so 2 to 30 m data is usually the right trade-off between coverage and cost. It's enough to track erosion, shoreline change, and burn-scar recovery across a full season without tasking sub-meter imagery over every square kilometer.

Disaster response

Speed matters more than pixel size in the first hours after a flood, wildfire, or earthquake. SAR imagery at 1 to 3 m images through cloud and smoke the moment it's tasked, while 30 cm to 2 m optical imagery follows once skies clear, confirming the extent and severity of the damage building by building.

Urban and government planning

Tracking how a city or region changes over years calls for consistent, repeatable coverage more than extreme sharpness. 2 to 10 m data is enough to classify land use, watch a city's boundary expand, and keep zoning maps current, at a cost that supports monitoring an entire municipality rather than a single site.

Frequently asked questions

What is satellite image resolution?

Satellite image resolution, also called spatial resolution or ground sample distance (GSD), is the size of ground area that one pixel represents. A 30 cm resolution image means each pixel covers a 30 by 30 centimeter square on the ground; a 10 m resolution image means each pixel covers a 10 by 10 meter square.

What is considered high resolution for satellite imagery?

In commercial satellite imagery, anything from 25 cm to 1 m per pixel is generally considered high or very high resolution. Below 25 cm is super high or ultra-high resolution, while 2 to 30 m is considered moderate to coarse resolution, still useful for regional and environmental analysis.

What is sub-meter resolution satellite imagery?

Sub-meter resolution means each pixel covers less than 1 meter of ground. It spans the super high (25-30 cm), very high (31-50 cm), and high (51-80 cm) tiers. Below that 1 meter line, individual vehicles, road markings, and building foundations are still visible; above it, features blur into approximate shapes and eventually disappear entirely.

What resolution of satellite imagery do I need for my project?

It depends on your smallest target feature. Vehicles, building foundations, and infrastructure detail need 25 to 50 cm. Crop fields, forest blocks, and land-use boundaries are well served by 2 to 10 m. Mineral alteration mapping and regional water basins are best matched to 30 m hyperspectral data. Buying finer resolution than the task requires just adds cost without adding useful information.

What is the difference between spatial and spectral resolution?

Spatial resolution is how much ground area one pixel covers, which determines visual sharpness. Spectral resolution is how many separate wavelength bands the sensor records, which determines how well it can distinguish materials by their light signature. A sensor can have a coarse pixel size but hundreds of spectral bands, as hyperspectral satellites do.

What is the difference between spatial and temporal resolution?

Spatial resolution refers to the level of detail visible in a single image. Temporal resolution refers to how often a satellite revisits the same location. The two typically trade off: a narrow, high-resolution sensor covers less ground per pass and takes longer to revisit a given area than a coarser, wide-swath sensor.

Why does higher satellite resolution cost more?

Finer resolution sensors cover a smaller ground footprint per pass, need larger optics or antennas, and produce far more data per square kilometer to store and process. That combination of smaller coverage and higher data volume is why 25 to 30 cm imagery costs roughly 20 times more per square kilometer than 2 m imagery.

Can satellite imagery see through clouds at any resolution?

Optical satellites cannot, regardless of resolution tier, since they depend on reflected sunlight that cloud cover blocks. Synthetic Aperture Radar (SAR) satellites, such as the GF-3 constellation, generate their own microwave signal and can image through cloud, fog, and darkness at resolutions from 1 to 500 meters.

What is the difference between archive and new tasking imagery?

Archive imagery is previously collected data, cheapest when it is 90 or more days old, and is the fastest and most affordable option for historical analysis. New tasking schedules a fresh satellite pass over your exact area and date, required when current ground conditions or a specific capture window matter, and typically costs 40 to 60% more per square kilometer than archive data.

What is radiometric resolution in satellite imagery?

Radiometric resolution is the number of brightness levels a sensor can record per pixel, measured in bits. An 8-bit sensor records 256 shades per band, while an 11-bit or 16-bit sensor records thousands, which allows it to distinguish subtle differences, such as slight variations in water turbidity or early vegetation stress, that a lower bit-depth sensor would flatten into a single shade.

Sources and further reading

• China Siwei and 21AT: SuperView Neo, SuperView-1/2, Beijing-3, and GF series satellite specifications
• CNSA and CRESDA: GF-5/5B, GF-6, GF-4, GF-3, and ZY-3 mission specifications
• ESA Copernicus: Sentinel-1 and Sentinel-2 mission specifications
• NASA/USGS: Landsat 4–9 mission history and archive access
• XRTech Group: satellite imagery pricing and collection speed tiers, 2026