Automatic character recognition system maps

Abstract. Based on the analysis of theoretical material pattern recognition theory, the authors of the algorithm and software was created by performing a search, identification and grouping of text and graphic symbols on maps to solve the problem of building geographic information systems areas. To find the optimum detection method, a comparison was made of three basic methods: comparison matrix, selection of features and intelligent method of artificial neural networks; developed algorithms for pre-processing and post-processing. The best quality has shown the method of artificial neural networks, and it is recommended for character recognition in solving the problem.