الفهرس 
Fully unsupervised hyperspectral image analysis
SummaryOnly 14 pages are availabe for public view
 |  | from | 132 |  |  |
| | | from | 132 | | |
Abstract
Spectroscopy is the study of light as a function of wavelength that has been emitted, reflected or scattered from a solid, liquid, or gas. Each material has its own spectral signature, and hence can be identified using spectral analysis. Hyperspectral imaging (HSI) - also called Imaging spectroscopy- sensors observe hundreds or thousands of contiguous spectral bands as well as spatial locality. A hyperspectral image cube (two spatial dimensions and the third is the wavelength) contains a large amount of information about the imaged scenario. Thus the automated analysis of such image cubes is an important asset. Spatial analysis in HSI is rather difficult due to the fact that HSI taken by a satellite or an airborne camera has low ground sampling distance (GSD). This means that many targets of interest can be located within one pixel. Another related problem is the availability of materials in nature as mixtures. As a result, spectral analysis is of great interest specially sub-pixel detection algorithms and spectral unmixing