基于图像的电动操作手柄装配缺陷视觉检测系统

Image-Based Visual Detection System for Assembly Defects of Electric Operating Handles

摘要: 电动操作手柄是断路器的关键操作部件，直接决定电路通断控制的可靠性与设备运行安全性。存在装配缺陷的手柄会大幅提升断路器故障概率、增加运维风险，因此手柄装配质量检测是生产流程中的核心环节。传统人工目视检测效率低、易漏检误检，难以满足大批量、高精度、连续化生产的质检需求。为解决上述问题，本文以断路器电动操作手柄为研究对象，采用YOLOv8-OBB旋转目标检测算法构建机器视觉装配缺陷检测系统，实现标签偏移、二维码失效、限位片缺失、螺钉漏装等典型装配缺陷的精准识别。实验结果表明，本系统对各类缺陷部件检测的mAP@0.5达0.99以上，检测速度为30 FPS，精度与实时性满足工业现场要求。同时，系统采用轻量化设计并完成硬件部署，具备低成本、易落地的优势，可为电力装备装配自动化质检提供稳定可靠的技术方案。

Abstract: The electric operating handle is a key component of circuit breakers, which directly determines the reliability of circuit on-off control and the safety of equipment operation. Handles with assembly defects greatly increase the failure probability of circuit breakers and raise operational and maintenance risks. Therefore, the assembly quality inspection of handles is a core part in the production process. Traditional manual visual inspection suffers from low efficiency, missed detection and false detection, which can hardly meet the requirements of high-volume, high-precision and continuous production. To address these problems, this paper takes the electric operating handle of circuit breakers as the research object and constructs a machine vision-based assembly defect detection system using the YOLOv8-OBB rotated object detection algorithm. The system realizes accurate identification of typical assembly defects such as label offset, invalid two-dimensional code, limit plate missing and screw omission. Experimental results show that the mAP@0.5 of the proposed system for all kinds of components exceeds 0.99, and the detection speed reaches 30 FPS. The accuracy and real-time performance meet industrial field requirements. Meanwhile, the system adopts a lightweight design and completes hardware deployment, featuring low cost and easy implementation. It can provide a stable and reliable technical solution for automatic quality inspection in power equipment assembly.