ABSTRACT
This chapter is concerned with multi-spectral images, i.e., with several channels that can correspond to various physical measurements (colors, wavelengths, features, etc.). Multispectral imaging allows the extraction of additional information that the human eye cannot necessarily capture. It permits the exploitation of redundancies and complementarities between the acquisition channels, and in addition to the extraction of relevant information. The analysis of these images can lead to a reduced and synthetic representation of the information contained in the whole set of channels. Several reference algorithms are presented and analyzed for dimension reduction of multi-spectral images, as well as color space transformations or multi-spectral edge detection.
