1. Hyperspectral imaging applications

Hyperspectral imaging enables us to obtain a spectrum for each point in the image, but it is clear that acquiring the spectrum of each pixel is only useful if we are able, on the one hand, to relate the spectral information to the physical characteristics of the object observed (with the exception of purely statistical processing, such as anomaly detection) and, on the other, to process the very large quantity of data generated: an image of 1024 × 1024 pixels, over 256 bands, coded in floating-point numbers, thus represents 1 GB of data. Increased computing capacity is largely responsible for recent developments in spectro-imaging. Provided these two conditions...