Computational diagnostics of objects that allow the heuristic strengthening of algorithms for solving reverse problems

• Published in: Vestnik Tomskogo gosudarstvennogo universiteta. Upravlenie, vychislitel'naya tekhnika i informatika no. 70; pp. 122 - 133
• Main Authors: Zerkal, Sergey M., Peshkov, Alexander V.
• Format: Journal Article
• Language: English; Russian
• Published: 01.03.2025
• ISSN: 1998-8605; 2311-2085
• DOI: 10.17223/19988605/70/12

The task of diagnosing industrial products for the presence of thin-length defects such as cracks and abrasions is considered. This problem is solved in the conditions of tomographic reconstruction and belongs to the class of reverse tasks. The classic algorithm of computed tomography has a high complexity in solving this problem, and in the case of a small thickness of the defect, it may not restore it at all. In this paper, software has been developed to solve a special problem of flaw detection based on a modified algorithm that uses a priori information about the reference of the product as a heuristic amplification. The results of computational experiments demonstrate the effec-tiveness of the proposed approach in solving the problem of detecting cracks in the object under study and allow us to conclude that its prospects in the tasks of industrial flaw detector.