摘要
手机取卡针在生产过程中常出现表面划伤、磨伤、异色等缺陷,这些缺陷严重影响产品的质量与交付。针对现有取卡针缺陷检测方法的局限性,该研究构建了基于YOLOv8的取卡针缺陷检测模型,对3种常见的取卡针表面缺陷进行学习、训练和验证。通过数据集构建、模型训练和模型评测3个步骤,实现了对手机取卡针表面缺陷的自动检测。实验结果表明,该模型在多类别缺陷检测中的平均精确度达到了98.45%,其中划伤、磨伤和异色缺陷的识别精确度分别达到了98.89%、97.87%和96.98%,验证了模型的有效性。此外,YOLOv8模型具有内存占用小、检测速度快等优势,显著提升了模型的工程适用性。
The surface scratches,abrasions,discoloration,and other defects often occur in the production process of the mobile phone card removal pin,which seriously affect the quality and delivery of products.To address the limitations of the existing card removal pin defect detection methods,this research constructs a card removal pin defect detection model based on YOLOv8 for learning,training,and verifying three common card removal pin surface defects.Through the three steps of dataset construction,model training,and model evaluation,the automatic detection of the surface defects of the mobile phone card removal pin is realized.The experimental results show that the average accuracy of the model in multi-category defect detection reaches 98.45%,and the recognition accuracies of scratches,abrasions,and discoloration defects reach 98.89%,97.87%,and 96.98%,respectively,verifying the effectiveness of the model.In addition,the YOLOv8 model has the advantages of small memory usage and fast detection speed,which significantly improves the engineering applicability of the model.
作者
陈光伟
江永春
CHEN Guangwei;JIANG Yongchun(Qingdao University,Qingdao 266071,China)
出处
《现代信息科技》
2025年第4期58-63,共6页
Modern Information Technology
关键词
YOLOv8
手机取卡针
缺陷检测
检测速度
YOLOv8
mobile phone card removal pin
defect detection
detection speed
