It is most important that all spatial data in GIS are located with respect to a common frame of reference. For most GIS this frame is provided by one of a limited number of geodetic coordinate systems.

Georeferencing is the process of establishing the relationship between an image (row, column) coordinate system, called image space, and a map (x,y) coordinate system, called map space. A similar transformation process occurs when establishing the relationship of a vector coverage in one coordinate system (i.e., digitizer units) to another map coordinate system. The form of the input data (a vector model, an image, a cell-based raster model) does not matter; if analysis is to take place with data on different data models, the data must be registered to a Cartesian coordinate system.

Georeferencing images is an important step for overlay techniques and to digitize maps after published paper maps. To be integrated into GIS, a non-georeferenced image (e.g. satellite image or scanned paper map image) must be converted from image space to map space. The image needs to be registered to a set of coordinates that are in map space. The parameters necessary to register can be determined in two ways 1) by extracting them from the information that accompany the image, or 2) by using georeferencing tools to transform the image. Depending on the means of collection, images can exhibit a variety of spectral and geometric distortions. Image-processing techniques like rectifying and rubber sheeting can be used to correct many types of distortions. Once registered and rectified, the image or grid can be used in conjunction with other georeferenced vector or raster models or images