Contour Method of Tomographic Scanning with Identification of Defects Using Computer Vision

Studying large objects is one of the most common problems of X-ray tomographic scanning, the solution of which requires the use of more powerful radiation sources, complex expensive mechatronics, and large-sized detector devices, which undoubtedly leads to a multiple increase in the cost of the X-ray unit itself. This article presents one of the possible methods for solving this problem, the essence of which is to scan objects along their contour. This approach can greatly reduce the cost of components of the X-ray unit. At the same time, the approach has a significant limitation: the presence of a large number of artifacts that do not allow detecting defects with sufficient reliability. This problem is proposed to be solved using machine learning.