The module SEGCLASS classifies the imagery using a majority rule algorithm to assign each
segment to the majority class from the reference image. SEGCLASS can improve the accuracy
of a pixel-based classification and produce a smoother map-like classification result while
preserving the boundaries between the segments.