基于改进YOLOv5算法的齿轮表面缺陷检测方法  

Gear Surface Defect Detection Method Based on Improved YOLOv5

在线阅读下载全文

作  者:马天龙 李学伟[1] 丁峰 潘广浩 郑光明[1] MA Tianlong;LI Xuewei;DING Feng;PAN Guanghao;ZHENG Guangming(School of Mechanical Engineering,Shandong University of Technology,Zibo 255000,China;Yishui County Inspection and Testing Center,Linyi 276499,China)

机构地区:[1]山东理工大学机械工程学院,淄博255000 [2]沂水县检验检测中心,临沂276499

出  处:《组合机床与自动化加工技术》2025年第4期152-156,160,共6页Modular Machine Tool & Automatic Manufacturing Technique

基  金:国家自然科学基金项目(51505265);山东省科技型中小企业创新能力提升项目(2022TSGC2260)。

摘  要:为解决传统齿轮缺陷检测易受环境影响、检测多特征缺陷效率低等问题,提出基于改进YOLOv5算法的齿轮表面缺陷检测方法。首先,为充分融合齿轮缺陷特征,对YOLOv5颈部网络的PANet结构进行改进,将其替换为BiFPN模块;然后,在骨干网末端加入ECA注意力机制,使模型更加关注缺陷区域的特征提取;最后,为了更合理地表示真实框和预测框之间的损失,采用SIOU作为损失函数。结果表明,改进后的YOLOv5模型的mAP达到87.4%,与原来的YOLOv5模型相比,mAP提高了12.1%,FPS达到了85.2 f/s。与Faster R-CNN、YOLOv4、YOLOv5等其他算法相比,改进后的算法在检测精度和模型大小以及速度方面均具有明显的优势,能够更加准确智能地检测齿轮表面缺陷。To solve the problems that the traditional gear defect detection is easy to be affected by the environment and low efficiency in detecting multi-feature defects,a gear surface defect detection method based on the improved YOLOv5 algorithm was proposed.First,in order to fully integrate gear defect features,the PANet structure of YOLOv5 neck network was improved and replaced with BiFPN module.Then,ECA attention mechanism is added to the end of the backbone network to make the model pay more attention to the feature extraction of the defect region.Finally,in order to more reasonably represent the loss between the real box and the predicted box,SIOU is used as the loss function.The results show that the mAP of the improved YOLOv5 model reaches 87.4%,which is 12.1%higher than that of the original YOLOv5 model,and the FPS reaches 85.2 f/s.Compared with other algorithms such as Faster R-CNN,YOLOv4 and YOLOv5,the improved algorithm has obvious advantages in terms of detection accuracy,model size and speed,and can detect gear surface defects more accurately and intelligently.

关 键 词:齿轮 缺陷检测 YOLOv5 注意力机制 损失函数 

分 类 号:TH16[机械工程—机械制造及自动化] TG66[金属学及工艺—金属切削加工及机床]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象