The raster data model represents the world as a regular grid of cells, each holding a value — an elevation, a temperature, a reflectance, a land-cover class. Satellite images, aerial photos, digital elevation models and modelled surfaces are all rasters. Where vector data describes discrete things, raster data describes continuous fields, making it the native language of remote sensing and environmental modelling.
Cells, resolution and extent
A raster is defined by its origin, cell size and row-column dimensions. Cell size — the ground distance one pixel covers — is its spatial resolution: a 10 m Sentinel-2 pixel averages everything within a 10 × 10 m square. Choosing resolution is always a trade between detail, coverage and data volume, since halving cell size quadruples storage.
Because every cell location is implicit in the grid, rasters need no per-feature coordinates. That makes them compact for full-coverage phenomena and extremely fast for algebraic operations across whole areas.
Bands and multi-layer rasters
Imagery rasters usually carry multiple bands — co-registered grids recording different wavelengths. A natural-colour image has red, green and blue bands; multispectral sensors add near-infrared and shortwave infrared; hyperspectral sensors carry hundreds of narrow bands. Band combinations and ratios, such as NDVI, unlock information invisible in any single band.
Time can be stacked the same way: a cube of monthly NDVI rasters becomes a per-pixel time series, the foundation of crop monitoring and change detection.
Map algebra and raster analysis
Raster analysis is grid arithmetic — map algebra. Slope from elevation, flood depth from water surface minus terrain, suitability from weighted layer sums: each is a cell-by-cell computation across millions of pixels. Reclassification, focal statistics, zonal summaries and cost surfaces round out the toolkit.
Modern formats such as Cloud-Optimised GeoTIFF (COG) let analysts stream exactly the pixels they need from web storage, which has quietly revolutionised how large raster archives are used.
Frequently asked questions
What is the main difference between raster and vector data?
Raster represents space as a continuous grid of valued cells, ideal for imagery and surfaces; vector represents discrete features with exact geometries and attributes, ideal for parcels, assets and networks. Most real projects use both.
What does spatial resolution mean in a raster?
It is the ground size of one cell. A 30 cm orthophoto resolves individual road markings; a 250 m climate grid smooths whole neighbourhoods into one value. Resolution should match the smallest object your analysis must distinguish.
Related topics
Need this done professionally?
GLOBEIR delivers this work for government and enterprise clients.