Abstract
In order to shorten the onboarding time of the new hires that need to be trained for working with pneumatic control systems, it is possible to develop an augmented reality (AR) application that would provide additional digital information about the components contained within the pneumatic control circuits. As the first step, this paper shows the first phase of work, a desktop application that was developed in order to enable the successful scanning of pneumatic control circuits and clear recognition of pneumatic symbols. This would enable the identification of pneumatic components used in a given system. A deep learning model was trained using the YOLOv5 algorithm to facilitate symbol recognition. This algorithm provided accurate and reliable results during pneumatic control circuit scanning.

This publication has 12 references indexed in Scilit: