Intelligent processing and analysis of groups of multispectral images

In this paper is offered one new approach for efficient processing and analysis of groups of multispectral images of same objects. It comprises several tools: the Inverse pyramid decomposition for still images; the invariant object representation with Modified Mellin-Fourier transform, and the hierarchical search in image databases, for which the invariant representation is used. The new approach permits to define large number of parameters, which are used for object analysis and evaluation. When combined with the KASER expert system, this approach turns into a flexible tool for analysis of multispectral images of the same object.