基于改进YOLOx网络的金属齿轮表面缺陷检测方法  

Gear Surface Defect Detection Method Based on Improved YOLOx Network

作　　者：张曙文 钟振宇 朱大虎 Zhang Shuwen;Zhong Zhenyu;Zhu Dahu(Hubei Key Laboratory of Advanced Technology for Automotive Components,School of Automotive Engineering,Wuhan University of Technology,Wuhan 430070,Hubei,China;Hubei Collaborative Innovation Center for Automotive Components Technology,School of Automotive Engineering,Wuhan University of Technology,Wuhan 430070,Hubei,China)

机构地区：[1]武汉理工大学汽车工程学院现代汽车零部件技术湖北省重点实验室,湖北武汉430070 [2]武汉理工大学汽车工程学院汽车零部件技术湖北省协同创新中心,湖北武汉430070

出　　处：《激光与光电子学进展》2023年第22期272-282,共11页Laser & Optoelectronics Progress

基　　金：国家自然科学基金(51975443)。

摘　　要：针对工业干扰环境下金属齿轮表面缺陷自动化检测容易出现误检和漏检的问题,提出一种改进的YOLOx算法。首先,通过adaptively spatial feature fusion(ASFF)充分利用不同尺度下缺陷和干扰项的特征之间的差异,提高模型的抗干扰能力;接着,通过efficient channel attention(ECA)模块,增加网络的特征提取能力;最后,修改置信度损失函数为Varifocal损失函数,减少困难样本对网络的干扰。实验结果表明,改进的YOLOx网络与原版相比在召回率、准确率和平均精度均值上分别提升6.1个百分点、4.6个百分点和9.4个百分点。Herein,an improved YOLOx algorithm is proposed to address the challenges concerning false and missing detection of metal gear surface defects in an industrial interference environment.First,by utilizing the adaptive spatial feature fusion(ASFF)to fully utilize the differences between the features of defects and interference items at different scales,the model’s anti-interference ability is improved.Second,through the effective channel attention(ECA)module,the network’s feature extraction capability is increased.Finally,the confidence loss function is modified to the Varifocal loss function,which reduces the interference of complex samples in the network.Experimental results indicate that the improved YOLOx network outperforms the original network.Particularly,the recall rate,accuracy,and mean average precision indexes of the improved YOLOx network are improved by 6.1,4.6,and 9.4 percentage points,respectively,as compared with the original network.

关 键 词：机器视觉 齿轮缺陷检测 YOLOx算法 注意力机制 特征融合 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]