The spectral properties
of the training sets along with previous knowledge, and data
from field observations were all combined to perform the supervised
classification (Veerendra and Latha, 2014). Analysis was
performed using the maximum likelihood classification and a total
of 12 land use classes were obtained.