Potential of hyperspectral remote sensing for analyzing the urban environment

Kurzfassung

In contrast to widely used multispectral data, hyperspectral imagery resolves material-specific spectral reflection and absorption features making them especially suitable for detailed and comprehensive mapping of urban surface materials. However, this requires development of automated methods for efficient information extraction that take into account the special conditions in urban areas characterized by a big variety of materials and a large heterogeneity of small-sized urban structures. This chapter deals with the current state of methodological developments in urban hyperspectral remote sensing emphasizing the research of the authors towards the development of an automated system for comprehensive mapping of urban surface materials. This includes field- and image-based spectral investigations aiming at the automated derivation of robust quantitative spectral features. They serve as input information for the developed multi-step processing system allowing detailed mapping of urban surface materials. In this context, the iterative procedure is capable of analyzing urban structures at a sub-pixel level and provides area-wide information about the fractional coverage of surface materials for each pixel. Thus, the obtained results are characterized by a new level of thematic and spatial detail which significantly increases their suitability for subsequent modeling, evaluation and monitoring of the urban environment.