Affine invariant hyperspectral image descriptors based upon harmonic analysis

摘要: This chapter focuses on the problem of recovering a hyperspectral image descriptor based upon harmonic analysis. It departs from use integral transforms to model images in terms probability distributions. provides link between analysis and affine geometric transformations object surface planes scene. Moreover, permits study these descriptors context Hilbert spaces. This, turn, connection functional capture spectral cross-correlation bands for generation with high energy compaction ratio. Thus, can be computed orthogonal bases capable capturing space wavelength correlation spectra imagery under study. We illustrate utility our purpose recognition dataset real-world objects compare results those yielded using an alternative.