Spectral Mixture Analysis
This chapter reviews the distinct field of spectral mixture analysis of remote sensing images. The problem consists on extracting spectrally pure pixels directly from the images. This extracted endmembers are very useful for image analysis and characterization.
As we have seen throughout the book and summarized in Chapter 1, hyperspectral sensors acquire the electromagnetic energy in a high number of spectral channels under an instantaneous field of view. These rich datacubes allow us identifying materials in the scenes with high detail. However, high spectral resolution is achieved at the price of spatial resolution due to technical constraints, and the resulting pixels are spatially coarse. ...