The Method of One-Dimensional Structural Approximation of Images of Cellular Surfaces of Industrial Products in the Task of Input Metrological Control

Abstract

The work is devoted to the development of structural approximation algorithms for one-dimensional analysis of raster images of complex surfaces for identifying the spatial coordinates of elements in a surface image. In particular, such surfaces include cellular surfaces of hulls and large shell structures of industrial products. The one-dimensionality of the analysis achieved by synthesizing one-dimensional structural approximation models of all horizontal and vertical lines of the image. The algorithms based on a general formal scheme of structural-approximation synthesis of one-dimensional images of objects with a regular structure, the use of which in various applications considered by the authors earlier in a number of publications. The proposed algorithms allow you to perform spatial identification of cells in the image from the camera. They can be used as part of software for metrological control of industrial products at the stage of preliminary input control. At such a control, gross defects are revealed, leading to disruption of the structure of the cellular surface specified in the design documentation.