A software package for tomographic diagnostics of industrial products using artificial intelligence tools

Computed tomography (CT) is becoming increasingly important in the field of non-destructive testing due to its ability to form three-dimensional images of the internal structures of objects. This allows for detailed analysis of complex geometric shapes and the detection of defects that may be inaccessible to traditional methods. CT provides high resolution and accuracy, which is critical for assessing the condition of products in industries such as aviation, automotive, and medicine. In addition, automation of CT processes and integration with quality management systems help to increase inspection efficiency and reduce costs. Thus, CT is a powerful tool for ensuring the reliability and safety of products. This article presents the results of developing a prototype software package for detecting small defects in industrial products using a tomographic approach. The core of the computing unit contains a previously developed computational-heuristic algorithm that allows detecting small-thickness defects in the product under study based on a priori information about the reference sample of this product. The developed software system is designed on the basis of a microservice architecture, which allows dividing the monolithic application into independently operating modules. The presented results demonstrate a significant increase in system performance compared to the previous implementation due to the use of a multi-threaded approach. The paper also describes an additional module that uses machine learning to determine the type of defect in an object. The results of the defect classification module demonstrate high accuracy rates and indicate the promise of using artificial intelligence in solving computed tomography problems. The developed prototype software system can be further developed and used to solve real problems in the field of industrial flaw detection.