الفهرس 
Comparison and Enhancement of Hyperspectral Unmixing Algorithms and Techniques
AbstractOnly 14 pages are availabe for public view
 |  | from | 94 |  |  |
| | | from | 94 | | |
Abstract
Hyperspectral imagery is a multispectral imagery with a massive amount of spectral information. The enhancement in the spatial dimension of the hyperspectral imagery, does not, always, match the enhanced spectral dimension. Therefore, with a low spatial resolution, a pixel may contain a mixture of materials which, in its turn, needs unmixing algorithm to extract the components of such pixels. The huge amount of spectral data, in hyperspectral imagery, makes it possible to dissolve each pixel, spectrally, to its constituents. To achieve this target, some matching methods, unmixing algorithms, must be applied between each pixel and the a priori well-known spectra that identifies materials in that pixel. Endmember Extraction Algorithms (EEAs) try to get all pure signatures from the scene in order that Hyperspectral Unmixing (HU) techniques use them as a priori signatures to unmix each pixel. In hyperspectral imagery, the role of endmember extraction lies in extracting distinct spectral signature, endmembers, from a hyperspectral image which is considered as the primary input for unsupervised hyperspectral unmixing to generate the abundance fractions for every pixel in hyperspectral data