This article is devoted to the applied aspects of a method of transforming images that uses the principle of similarity to a standard. The method is based on software that makes it possible to carry out the transformations either automatically or with the interactive participation of the user. The method is suitable for processing multichannel (multizonal, hyperspectral) images. The processing has the goal of making it possible to distinguish information that is useful for a given problem, condensing the initial data. This paper presents examples of the processing of hyperspectral images. Another area in which the method of transformation by the principle of similarity to a standard can be used is automatic coordinate registration of images of the same scene obtained under various conditions (at different times, in different channels).
© 2010 Optical Society of AmericaPDF Article