Raster data model

From wiki.gis.com
Jump to: navigation, search

Raster data models present information through a grid of cells.

Raster grids are usually made up of square or rectangular cells. Unlike vector data models, which show geographic data through points, lines, or polygons, raster data is displays one value in each cell. This value can be interpreted to mean several different things. Changing the cell size of Raster data can make the projected image more or less blurry, the . Raster data is especially effective at representing data that is continuous, such as elevation, precipitation, aspect, and slope.[1] All forms of digital imagery whether it is from low altitude UAV's, suborbital, or satellite is displayed in a raster cell grid. Lidar and radar imaging is done in raster. Raster data derived from lidar data is particularly useful in building digital elevation models.

Structure of Raster Data

The raster logical model represents a single geographic phenomenon (usually, but not always a field) as a two-dimensional array of samples, usually at regular spacing in both the x and y directions. Each sample having a single value. The meaning of the sample measurement depends on which of two conceptual models the measurement framework may use. In a lattice, each value represents a measurement at a single point. In a grid, each value represents a summary of the values within a square, such as mean temperature. Each horizontal array of values (having the same y value) is called a row, and each vertical array (having the same x value) is called a column. Each sample location (whether point or square) is called a cell, or pixel if the raster is an image.

In most raster data structures, the array is stored as an ordered list of values, either going across the cells in a row then going to the beginning of the next row (row-major) or going down the cells in a column then going to the beginning of the next column (column-major). A header is included at the beginning of the file, providing at least the basic metadata necessary to arrange the list as a two-dimensional grid in the proper spatial location:

  • the x,y coordinate of the sample in the first row and column (usually either the Northwest or Southwest corner)
  • the dimensions or spacing of each cell, known as the cell size or resolution. Traditionally, the cell size has been required to be the same in the x and y directions (i.e., the cells are square), but some modern structures allow this to vary.
  • the number of rows and columns. Technically, only one of these is required to know how to arrange the list as an array (depending on whether it is row-major or column-major), but both are usually included.
  • the data type of each value (e.g., short integer, long integer, double-precision floating point, ASCII character), specifically the number of bytes needed for each value.

Other helpful metadata is often included in the header, such as the coordinate system or the units of measure in which the values are given.

On top of this core data model, GIS and image file formats may add a number of extensions:

  • Compression. For analytical GIS data, lossless compression algorithms (such as LZW, RLE, or PNG) are required, while images can be represented by lossy algorithms (such as JPEG).
  • Mosaics. The Esri Geodatabase format, as well as other modern GIS formats, allows for multiple rasters (usually remote sensing imagery) to be collected and represented to users as a single seamless image, even though the individual rasters are still stored separately.
  • Scale Hierarchies. To improve drawing performance at smaller scales, many formats allow users to resample the original raster at a series of coarser resolutions, typically called a pyramid file. These can then be read directly when displaying at a corresponding scale rather than resampling at display time.

Advantages of Raster Data

One of the great advantages of Raster data is the ability to portray continuous data which other wise could not be displayed by points, lines, and polygons.

Limitations of Raster Data

Most of the output maps for raster data could not be used by cartographers because of it's inaccuracy. [2] The image of a non-contiguous maps will not be as accurate as vector data map would be due to the distortion created by the different cell sizes


  1. Paul Bolstad (2008). GIS Fundamentals: A First Text on Geographic Information Systems. Eider Press. ISBN 978-0-9717647-2-9. 
  2. Buckey, David. "Vector and Raster- Advantage and Disadvantages." Bgis Introduction to GIS. N.p., n.d. Web. 16 Jan. 2013.