A Study of Defective Wiring Detection Platform Based on Autonomous Mobile Robot in Small-Scale AI Smart Factory Environment

Abstract

This study proposes an automated system that combines a YOLOv4 based vision recognition technique with a robotic platform to efficiently detect wiring defects occurring in process systems using a self-built small-scale smart factory environment. Conventional manual inspection methods require significant time and cost, and their detection accuracy varies depending on the skill level of the operator. To address these limitations, the proposed system employs a mobile robot equipped with a camera that patrols production facilities, capturing real-time images of wiring conditions. Subsequently, the YOLOv4 algorithm is used to rapidly detect defects such as disconnections and insulation damage. In addition, a lightweight model structure is used to ensure real-time processing performance suitable for operation even in small-scale factories. Experimental results demonstrate that the proposed system achieves a detection accuracy of approximately 92.1% and a processing speed of over 24 FPS, indicating its practical applicability and effectiveness in real-world environments. This study can contribute to enhancing the level of quality control automation in smart factories and reducing reliance on human labor.