128
4: Stefan A. Robila
a
c
Fig. 4.7. One band of the AVIRIS image with four locations indicated for further analysis of
pixel vectors (a) road, (b) soybean, (c) grass, (d) hay
Background
Alfalfa
• Com-notill
• Com-min
• Com
_ GrassIPasture
_ Grassffrees
_ Grass/pasture-mowed
_ Hay-windrowed
Oat
_ Soybeans-notill
_ Soybeans-min
_ Soybeans-clean
_ Wheal
_Woods
BuiJdings/GrassffreeIRoads
Stone/Steel Towers
Fig.4.8. Reference data (ground truth) corresponding to AVIRIS image in Fig. 4.7. For
a colored version of this figure, see the end of the book
a
b
c
d
Fig.4.9a-d. Resulting ICA components. Notice that the features correspond to information
related to various classes in the scene (such as a road, b soybean, c grass, d hay) are
concentrated in different components
4: Stefan A. Robila
a
c
Fig. 4.7. One band of the AVIRIS image with four locations indicated for further analysis of
pixel vectors (a) road, (b) soybean, (c) grass, (d) hay
Background
Alfalfa
• Com-notill
• Com-min
• Com
_ GrassIPasture
_ Grassffrees
_ Grass/pasture-mowed
_ Hay-windrowed
Oat
_ Soybeans-notill
_ Soybeans-min
_ Soybeans-clean
_ Wheal
_Woods
BuiJdings/GrassffreeIRoads
Stone/Steel Towers
Fig.4.8. Reference data (ground truth) corresponding to AVIRIS image in Fig. 4.7. For
a colored version of this figure, see the end of the book
a
b
c
d
Fig.4.9a-d. Resulting ICA components. Notice that the features correspond to information
related to various classes in the scene (such as a road, b soybean, c grass, d hay) are
concentrated in different components