This map illustrates the development of a land use/land cover
classification map based on digitizing recognized ground features in a natural
color aerial photograph. Classification was based on the standard USGS
Level I and Level II classification system. Even though schemes for
higher classification exist, they are typically used for specialized
projects. Here, students were to create their own Level III
classifications based on the features found in the study area. Following the classification, 35 sample
points (shown as small circles on the map) were selected as
ground truth locations for the classified areas. The sample points
were chosen based on a stratified random sampling scheme, where samples were
taken in each classification category. However, fewer samples were
taken in regions that exhibited little variability (e.g., marshes and near
shore coastal waters) while a larger number of samples were taken on the
land areas of interest. The locations were "truthed" using
Google's Street View feature. Overall classification accuracy was
then computed.
