Abstract
The ROI of the article relates to computer vision. Mathematical foundations and algorithms for preparing two-dimensional graphic codes for data decoding are presented. The preparation stages include the detection, localization and transformation of the code to a decoding-friendly form. The brightness oscillation function proposed in the paper serves to reliably detect codes in the image, the Hough transform is the basis for solving the localization problem, and the formulas for the inverse transformation of a square into an arbitrary quadrilateral allow you to transform localized quadrangular areas into squares with a horizontally oriented base. The work of algorithms on a set of test and real images is studied, the practically permissible level of distortion during the transformation of detected and localized two-dimensional codes into the original square form is shown.
Publisher
Keldysh Institute of Applied Mathematics