Raster Data
Definition
Raster data represent space as a grid of cells (pixels), each storing a value—reflectance, elevation, temperature, class code. Resolution controls detail; extent and alignment control coverage and interoperability. Rasters can be single-band or multi-band (multispectral). File formats (GeoTIFF/COG, NetCDF, HDF) store georeferencing, nodata, and metadata. Strengths include efficient storage for continuous fields and compatibility with image processing; limitations include stair-step edges and large volumes for high resolution. Understanding coordinate reference systems, compression, tiling, and pyramids is key for performance. Many analyses—terrain derivatives, indices, convolution—operate naturally on rasters.
Application
Remote sensing, climate modeling, DEMs for hydrology, hazard maps, and land-cover products all use rasters. Web maps serve raster tiles for imagery and hillshade. Scientists share gridded datasets for reproducible modeling.
FAQ
How do COGs improve sharing?
Cloud-Optimized GeoTIFFs allow HTTP range reads so clients fetch only needed tiles; they embed overviews for quick previews.
What determines appropriate resolution?
Match to phenomenon scale and decision needs. Finer isn’t always better—noise and storage grow while insight may not.
How to keep rasters aligned?
Use common origins, cell sizes, and CRS; snap to a grid before analysis to avoid resampling artifacts later.
Are rasters only for continuous data?
No—categorical rasters store classes; just use nearest-neighbor resampling and appropriate color tables.