Semi-trailer assembly line production control using computer vision

The paper explores the problem of using machine vision for quality management as part of the organizational system of a production workshop. It is proposed to solve the problem of the limited use of artificial neural networks in production by combining machine vision technologies and technologies for positioning objects in space. The chosen approach allows us to avoid the need to accumulate large volumes of training samples and limit the use of computer vision to specific tasks of identifying typical operations and various violations of the production process. The proposed approach was tested on a semi-trailer assembly line. Implemented functionality includes general employee control, security control, counting of manufactured products, personalized control of employee movement, as well as the process of collecting general analytics, which allows you to generate statistics for each post, section, block of workshops and for the entire enterprise as a whole. The results of the implementation of the proposed solutions at the semi-trailer assembly site confirmed the possibility of using machine vision systems in a modern manufacturing enterprise as part of the automation of production monitoring, quality control of processes and products and support for organizational decision-making support. Testing and implementation of the proposed solutions confirmed their practical usefulness for a manufacturing enterprise.