Hyperspectral Remote Sensing

Hyperspectral images provide a very detailed representation of the spectral reflectance of the target surface. Such data have been used in several applications such as land cover maps generation and agricultural monitoring and methodologies including classification and change detection where the dense sampling of the spectrum can be used to separate classes or changes associated to very similar, yet semantically different, spectral signatures which are undistinguishable in MS images. Indeed, compared to MS, sensors which sample spectrum at few broad bands, HS imagers perform a dense sampling of the spectrum across the whole spectral range of interest. However, the analysis of HS images requires to address several challenges related both to disturbances and distortion introduced during the acquisition and the processing of high dimensional data. While to this date HS images are still relatively less used than MS images, this will significantly change in the next years due to the current...