Skip to main content

Satellite imagery basics

Satellite imagery is most useful when the research needs repeat observations, wide coverage or spectral information, rather than the finest possible visual detail.

Four resolutions​

  • spatial — ground size represented by a pixel
  • temporal — how frequently the sensor can observe an area
  • spectral — number and width of wavelength bands
  • radiometric — sensitivity/bit depth used to record intensity

A sensor with coarser pixels can be better for monitoring if it returns every few days and has useful NIR/SWIR bands.

Optical​

Optical systems measure reflected sunlight. Common products include true colour and multispectral bands. Cloud, haze, sun angle and seasonal illumination matter.

Radar​

Synthetic aperture radar actively transmits microwave energy and records backscatter. It works at night and through cloud. Radar is useful for flood extent, deformation and surface/vegetation structure but requires different interpretation from photography.

Common open sources​

  • Sentinel-2 — multispectral optical data, strong for vegetation/water/land-cover monitoring
  • Landsat — long historical satellite record
  • Sentinel-1 — radar; useful where cloud blocks optical observations
  • NASA Earthdata — discovery/access to many Earth-observation missions

Commercial systems such as Planet, Maxar and Airbus can provide higher spatial resolution or revisit performance, but access/licensing varies and should be checked for each project.

Start with the research scale​

A 10 m pixel is excellent for catchment vegetation change and usually poor for identifying a small building. Do not resample a 10 m image to 1 m and treat it as new detail.

Last reviewed: 26 August 2026

Last updated on