Methods for Recognizing the Structure of Mounting Schemes in Railway Automation and Remote Control Systems

This paper is about the module for recognizing the structure of mounting technical documents for railway automation and telemechanics systems. The operation of the recognition module is divided into five stages: document type identification, recognition of basic elements, sorting and pattern recognition, template matching and missing elements identification, final validation and error detection. The algorithms for recognizing tabular structures of documents are considered in detail, in particular the algorithm for searching for basic elements and the algorithm for finding missing elements. Conventionally, mounting technical documents of railway automation and telemechanics systems can be divided into two basic types - tabular and schematic. In a document of a tabular type, the data are presented in the form of tables, in documents of a schematic type, the data are on a schematic image (for example, down terminal panels). To recognize the schematic type of installation diagrams of railway automation and telemechanics, a modification of the well-known algorithm for constructing the Zhang-Suen image skeleton, methods for preprocessing images are proposed, and the minimum required set of preprocessing algorithms is selected.