The Spectral Dimension in Urban Remote Sensing - Remote Sensing of Urban and Suburban Areas - page 51

Table 4.1 A land cover classiication scheme for urban environments
Level 1 Level 2 Level 3

Built up Buildings/roofs Composite shingle roof

Gravel roof
Metal roof
Asphalt roof
Tile roof
Tar roof
Wood shingle roof

Transportation areas Asphalt roads

Concrete roads
Gravel roads
Parking lots
Railroad

Other built surfaces Tennis court

Basketball field

Vegetation Green vegetation

Non-photosynthetic vegetation (NPV)

Non-urban bare surface Bare soil

Bare rock

Water bodies Natural/quasi-natural water bodies

Swimming pools

Fig. 4.4 Matrix of B-distance values for minimum and average spectral separability between different
urban land cover types derived from the Santa Barbara urban spectral library (Herold et al. 2004 )

or derived from sample areas in remote sensing image data. This spectral separability
technique calculates the spectral distance for all samples of a class to the ones of
another category. The result is an average and minimum separability score between
these two classes with higher B-distance values reflecting larger spectral separa-
tion. The resulting matrix in Fig. 4.4 containing the separability scores for several
urban land cover types are presented in Table 4.1 .