distinctive pattern of geomorphic zonation (i.e., forereef, reef crest, back reef, and
lagoon with pinnacles), which may be easily visible and occur at spatial scales of
tens to hundreds of meters. Investigating reef geomorphologic zonation
Landsat ETM+: Spatial resolution = 30 m, Accuracy = 64 %
IKONOS: Spatial resolution = 4 m, Accuracy = 81 %
Shiraho
Coral
Seagrass
Bare substrate (pavement)
Bare substrate (submerged sand)
Bare substrate (exposed sand)
Land
Ocean
Original image
Classified image
500 m
N
5 km
N
Fig. 3.1 Example of simple habitat classification using Landsat ETM+ and IKONOS images at
Shiraho Reef, Ishigaki Island, Japan (24°22
0 N, 124°15
0 E). Higher spatial resolution delivers
higher classification accuracy. To improve accuracy further, the habitat classified as ‘‘seagrass’’
in the ellipsoid indicated by the arrow on the IKONOS classified image can be corrected to
‘‘coral’’ using contextual editing (Andréfouët et al. 2003)
3 Multispectral Applications
57
lagoon with pinnacles), which may be easily visible and occur at spatial scales of
tens to hundreds of meters. Investigating reef geomorphologic zonation
Landsat ETM+: Spatial resolution = 30 m, Accuracy = 64 %
IKONOS: Spatial resolution = 4 m, Accuracy = 81 %
Shiraho
Coral
Seagrass
Bare substrate (pavement)
Bare substrate (submerged sand)
Bare substrate (exposed sand)
Land
Ocean
Original image
Classified image
500 m
N
5 km
N
Fig. 3.1 Example of simple habitat classification using Landsat ETM+ and IKONOS images at
Shiraho Reef, Ishigaki Island, Japan (24°22
0 N, 124°15
0 E). Higher spatial resolution delivers
higher classification accuracy. To improve accuracy further, the habitat classified as ‘‘seagrass’’
in the ellipsoid indicated by the arrow on the IKONOS classified image can be corrected to
‘‘coral’’ using contextual editing (Andréfouët et al. 2003)
3 Multispectral Applications
57