around a pixel to aid in its classification (Visual Learning
System Inc., 2004). Feature Analyst
®
has numerous options for
the selection of window shape and size. For this classification,
a Manhattan (diamond) shape with a 13 pixel-width was
selected. A minimum mapping unit (MMU) of 352 pixels or
approximately 0.04 hectares (0.1 acres) was used for all feature
classes. Utilizing the training set, each individual feature class
was classified. Then a “wall-to-wall” classification was created
using all the individual feature classes.