An Empirical Study on Process Management System using YOLO-Based Parts Recognition

Abstract The process of assembling the engines used in medium and large ships appears to be a continuous production process, but it is not the process through which numerous workers perform work with automated equipment or simple assembly work, but rather a process through which skilled workers, who can effectively perform various tasks, perform multiple aspects of the assembly process. Due to this characteristic, it was difficult to assess the work process rate during the engine assembly process through real-time analysis of the process because it was difficult to shorten the lead time or improve the process, and most of stages of the process depend on manual reports or operators’ work data input. Therefore, in this study, we developed and applied a system capable of recognizing the current process rate by identifying specific engine parts using YOLO.