Scale and Spatial Autocorrelation From A Remote Sensing Perspective.

One of the challenges for urban and regional planners and other users of remotely sensed imagery is how to select the appropriate data for a particular monitoring or mapping problem. In the past, the dearth of available imagery meant that the problem itself usually had to be adapted to fit the data, which was typically limited to either high spatial resolution film-based aerial imagery, or coarse-spatial resolution digital satellite imagery. Today, a vast range of aerial and satellite imagery is available (Kramer, 2002), opening a new range of potential scales of problems that can be investigated. However, these new options also place additional burdens on the remote sensing user, who, in selecting data, has to consider differences in spectral, temporal, radiometric, and spatial characteristics of the imagery. Spatial properties are particularly important, and the pixel size of current sensors varies over more than three orders of magnitude (from 0.6 m to 1 km and larger) (Kramer, 2002). [ABSTRACT FROM AUTHOR]