A data transformation method that rotates the axes of the input bands to
a new multivariate attribute space in which the axes are uncorrelated. The main reason to transform the
data in a principal component analysis is to compress data by eliminating redundancy.