We first built the endmember library for each year according to a fourendmember
model with shade endmember included into the
vegetationeimpervious surfaceesoil (VeIeS) model (Ridd, 1995).
Three methods were used to facilitate the identification of spectrally
“pure” pixels: pixel purity index (PPI) (Boardman, Kruse, and
Green, 1995), count-based endmember selection (CoB) (Roberts
et al., 2003), and endmember average root mean square error
(EAR) (Dennison & Roberts, 2003a).
The final set of end members
for the 2000 image consists of 4 spectra in the vegetation category,
15 distinct endmember spectra in the impervious surface category,
and 2 spectra in the soil category.