Glossary for Geospatial Science

boxcar classification or boxcar interpretation:�� The simplest form of automated image interpretation whereby three data ranges are selected for three coregistered images (like red, green, and blue).� The three data ranges define a boxcar shape if plotted on three dimensional perpendicular axes that represent possible data values in red, green, and blue.� The ranges are usually selected to represent the color variation in the three rasters for a feature of interest (like all the dark brown areas representing bare soil).

All of the data triplets in the three raster objects are tested to determine if they represent a cell inside the boxcar.� All the inside cells may then be displayed or otherwise recorded as �classified.�

Boxcar classification is not restricted to sets of three rasters.� It may be used with any number: using 2 rasters designs a 2-dimensional �boxcar� test in a 2-dimensional space, and using �n� rasters defines an n-dimensional �boxcar� test in n-dimensional space.