|
Using the MLRC landcover classification,
we mask out urban, residential, agricultural and water pixels from the
raw TM imagery. The remaining pixels are clustered based on their spectral
similarity. We then use the interpreted video points to query each layer
for the data value at that position. This creates an information matrix
that has vegetative alliance, cluster value, wetland type, wetland modifier
(eg. flooding regime), and soil type.
|
|
|