The purpose of this work is the application of neural network technologies in the production flow of metallurgical enterprises. The object of the study is a sheet-rolling shop. The subject of the research is the use of neural network technologies to track and identify sheets. In real workshop conditions, it is not possible to identify a sheet using text recognition technologies due to the impact of a huge number of disturbances, as can be seen from the figures in the article. The article presents the first important idea - to use the marking imprint as a master image for further sheet recognition. The second postulate of the article is training a neural network and identifying an object from its presence in the technological network. The third idea is to create a digital twin to simulate sheet movement. And, as a result, a neural network is configured to solve the above problems. The paper discusses in detail the use of the YOLOv8 model for object detection and convolutional neural networks for comparing pairs of images. The research results are expected to be used to improve the efficiency of monitoring and control of production processes at metallurgical enterprises.