改进的YOLOv5s布匹疵点检测算法研究  

Research on Improved YOLOv5s Fabric Defect Detection Algorithm

作　　者：徐春鸽 Xu Chunge(School of Data Science and Computer Science,Guangdong Peizheng College,Guangzhou,Guangdong 510830)

机构地区：[1]广东培正学院数据科学与计算机学院,广东广州510830

出　　处：《现代工程科技》2025年第2期1-4,共4页Modern Engineering Technology

摘　　要：在纺织行业中,布匹疵点将严重影响其经济价值,而目前纺织工厂的质检环节严重依赖人工质检。人工质检过程中会出现错检、漏检以及检测效率低下等问题。采用改进的YOLOv5s目标检测算法来实现布匹疵点检测。通过融入CBAM注意力机制、BiFPN模块,以提升对布匹缺陷的聚焦效果,有效地抑制了无效特征和噪声,提高了检测精度。实验结果表明,改进YOLOv5s-CB算法与原YOLOv5s算法相比,显著提高了布匹疵点的检测精度。In the textile industry,fabric defects will seriously affect its economic value.At present,the quality inspection of textile factories relies heavily on manual quality inspection.In the process of manual quality inspection,there will be problems such as false detection,missed detection,and low detection efficiency.The paper uses the improved YOLOv5s target detection algorithm to realize fabric defect detection.By integrating the CBAM attention mechanism and the BiFPN module to improve the focus of cloth defects,effectively suppress invalid features and noise,and improve detection accuracy.The experimental results show that the improved YOLOv5s-CB algorithm significantly improves the detection accuracy of fabric defects compared to the original YOLOv5s algorithm.

关 键 词：深度学习 布匹缺陷检测 YOLOv5s算法 YOLOv5s-CB算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]