Design of Machine Vision-Based Glass Defect Detection System

This paper addresses the challenges of time-consuming, labor-intensive, and inefficient glass defect detection in current manual visual inspection methods by proposing a machine vision-based glass defect detection system. The system comprises three components: a motion control system, an image acquisition system, and an auxiliary system. This study introduces a preliminary detection method combining synchronized strobe imaging and an enhanced YOLOv8-Glass network for glass defect identification. The overall system architecture is first defined, followed by the selection of key components and preliminary algorithm testing. Performance evaluations demonstrate that the proposed system achieves automatic defect recognition with high efficiency and intelligent operation during glass inspection, indicating its potential for industrial applications and confirming its superior detection capabilities.