Martin Herold, Margaret E. Gardner, Val Noronha and
Dar A. Roberts

Abstract | Full Paper (PDF, 2.61 MB)

Based on a comprehensive regional spectral library from spectrometer ground measurements and high-resolution hyperspectral AVIRIS data, this research investigates generic spectral characteristics, and related mapping capabilities and limitations for extracting roads of various types, delineating road centerlines, and surveying pavement quality within the urban environment. The methods are generally successful in identifying roads and road pavement types, but inaccuracies arise from spectral similarities between asphalt surfaces and specific roofing materials. Findings indicate that it is possible to spectrally characterize and estimate age of pavement and a few specific asphalt surface quality parameters (e.g. raveling). Finally, this research argues that the complex spectral characteristics of the urban environment result in distinct limitations in mapping urban areas using common multispectral sensor data. Sensors with an "optimized" set of narrow spectral bands would be preferred for prospective applications.

Martin Herold
Department of Geography
University of California Santa Barbara
Santa Barbara, CA 93106 USA
Tel: +1 805 893 4196
Email: martin@geog.ucsb.edu

back | top