Sydor A. Methods and Processors for Recognizing Multidimensional Images in the Hemming Space

The dissertation is devoted to the development of methods and processors for the recognition of multidimensional images in the Hemming space. The mathematical bases and prospects of development of theoretical foundations of image recognition methods in the Hemming space are described. Research of structural solutions of existing devices of correlation processing of signals as components of special processors for the recognition of multidimensional images. The method of estimating the Hemingway distance on the basis of the Euclidean distance and analyzes of various interrelation functions is developed. The method of processing signals of a two-dimensional Hemming space on the basis of transformations of the modular arithmetic of the Crescent's theoretic-numeric basis is developed. The method of determining the Hemming distance of images with different number of parameters is improved. The analysis of coding and definition of Hemingway distance for implementation of recognition of road signs of different categories is carried out. The developed structures and algorithms of the work of advanced special processors for estimating the Heming distance between the signals and implemented the microelectronic implementation of their components on the FPGA. The developed high-speed special processor for determining the Heming distance between the digitized signals in the theoretical-numerical basis of Haar-Krestenson, which, compared with known devices, is characterized by an increase of 2-3 orders of magnitude.