Raster and Vector - Rasters Defined

Concepts
   raster
   extent
   cell
   pixel
   cellsize
   resolution
   integer raster
   floating point raster
   value attribute table (VAT)

What is a raster?
A raster is a rectangular grid of cells which cover a geographic area.  The area a raster covers is its extent.  Each cell (or pixel) in a raster has a single numeric value assigned to it.  That value can represent any number of attributes, for example:
   - the elevation of the cell at its center point (this type of raster is known as a digital elevation model or DEM)
   - a code indicating the type of vegetation (or landcover) of the cell
   - the amount of light or energy captured by an airborne camera or satellite sensor (such as in aerial photos)
   - the color of a cell of a scanned-in map (such as USGS topo maps)
   - or any other type of value you could think of such as zoning class, development suitability, flood zone, etc.

A horizontal line of cells in a raster is known as a row and a vertical line is known as a column.  A raster is always defined by its starting point (lower left corner), its cellsize, and the number of rows and columns.

Cellsize/resolution
Each cell in a raster is a perfect square.  The length along one side is its cellsize (or resolution).  A 1-meter raster has cells which are 1m x 1m square, or 1 square meter.  A 10-meter raster has cells which are 10m x 10m square, or 100 square meters.  It is important to remember that cellsize is not a measure of area; it is the square root of the raster cell's area.



Integer/floating point raster
The value in the raster cell is always numeric.  Some rasters have this value as an integer, which is called an integer raster.  These are typically used with discrete data, such as landcover classes, from which there are a set number of values to choose.  Some rasters have this value as a decimal, which is called a floating point raster.  These are typically used with continuous data, such as elevation, from which there is an infinite range of values.



Value attribute table (VAT)
For integer rasters, the values of the cells are stored in a value attribute table which simply lists all possible values and the number of cells in the raster with that value.  VATs can have additional fields added them to explain in more detail what the value means.  For example, a landcover category of 201 does not mean anything to most people, but in the VAT a text field can be added indicating this is 'Abandoned Field'.



Floating point rasters typically do not have VATs built because continuous data would generate a huge number of records.  Only in very small floating point rasters can VATs be built.