Different processing techniques to analyze high resolution hyperspectral data of urban areas

Abstract

Large-scale urban green areas management tasks, like determining tree species and their degradation, require high geometrical accuracy, as well as efficient data extraction techniques. Are multi or hyperspectral image types suited for these tasks? To answer this question, the authors examined accuracy and extraction techniques for DAIS 7915, a hyperspectral sensor. Remote sensing software packages that were used were: PCI OrthoEngine 8.2 satellite edition, ERMapper, Envi and for the object oriented classification Definiens eCognition 2.0.