Spectrometry for urban area remote sensing—Development and analysis of a spectral library from 350 to 2400 nm

We investigate the spectral complexity and unique spectral characteristics of urban environments using a comprehensive regional field spectral library of more than 4500 individual spectra. Spectral properties of urban surface materials are presented and interpreted, and their separability is systematically analyzed using the Bhattacharyya distance (B-distance) as a quantitative measure of spectral discrimination. We find considerable spectral confusion between urban land cover types (i.e. specific roof and road types) but also show the potential of fine spectral-resolution remote sensing for detailed mapping of urban materials and their condition based on their spectral signal. An evaluation of the most suitable wavelengths for separation of urban land cover identified specific spectral features that provided the best separation. There is a strong indication that current multispectral systems, including IKONOS and LANDSAT ETM+, provide only marginal abilities to resolve these important features and are limited for urban land-cover mapping.