Korchyns'kyj V. Invariant geometrical models of identification and analysis of projective images

The projective images of flat objects including raster and fractal images. The new direction in geometrical modeling of the images of projective nature is created on the basis of the concept of geometrical object in many-dimensional space of cumulants of indicator functions. The theoretical bases of a direction are constructed: metrization of model space; conformity between the forms of the images and the points of this space. The applied methods are developed: image recognition invariant to the factors of their forming; restoration of mutual orientation of object and picture planes, topological characteristics of the images. The proposed methods of the image morphological analysis may be used for the automated processing of the videoinformation generated by rangefinder iconical systems, selection of objects under their geometrical forms, image compression.