基于改进YOLOv8的铝制品表面瑕疵检测研究  

Research on Surface Defect Detection of Aluminum Products Based on Improved YOLOv8

作　　者：陈晓桐 CHEN Xiao-tong(Fujian Business University,Fuzhou 350000,China)

机构地区：[1]福建商学院,福建福州350000

出　　处：《世界有色金属》2024年第23期28-30,共3页World Nonferrous Metals

基　　金：福建省自然科学基金资助项目(2020J01326)。

摘　　要：在工业应用,铝制品表面瑕疵直接影响产品质量和使用寿命。传统的瑕疵检测方法依赖人工检查,效率低且容易受到主观因素的影响。在YOLOv8算法的基础上引入了注意力机制和特征融合策略,提升模型对关键信息的捕捉能力和对复杂场景的适应能力。针对铝制品表面瑕疵检测的实际应用需求,对模型的检测速度和内存占用进行了优化。通过实验验证,该算法在多种瑕疵类型的检测中检测精度提升至94.2%,展现出较高的精度和召回率。In industrial applications,surface defects of aluminum products directly affect product quality and service life.Traditional defect detection methods rely on manual inspection,which is inefficient and susceptible to subjective factors.On the basis of the YOLOv8 algorithm,an attention mechanism and feature fusion strategy are introduced to improve the model's ability to capture key information and adapt to complex scenes.Aiming at the practical application requirements of surface defect detection of aluminum products,the detection speed and memory occupancy of the model are optimized.Experimental verification shows that the algorithm improves the detection accuracy to 94.2%in the detection of various types of defects,showing high accuracy and recall rate.

关 键 词：表面缺陷检测 YOLOv8m-CLS算法 注意力机制 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]