[1]于明源,周景亮,曾绍锋,等.基于YOLOv8改进的轴承表面缺陷检测方法[J].福建理工大学学报,2024,22(03):280-285.[doi:10.3969/j.issn.2097-3853.2024.03.011]
　YU Mingyuan,ZHOU Jingliang,ZENG Shaofeng,et al.Improved bearing surface defect detection method based on YOLOv8[J].Journal of Fujian University of Technology;,2024,22(03):280-285.[doi:10.3969/j.issn.2097-3853.2024.03.011]
点击复制

基于YOLOv8改进的轴承表面缺陷检测方法()
分享到：

《福建理工大学学报》[ISSN:2097-3853/CN:35-1351/Z]

卷:
第22卷
期数:
2024年03期
页码:
280-285
栏目:
出版日期:
2024-06-25

文章信息/Info

Title:
Improved bearing surface defect detection method based on YOLOv8
作者:
于明源; 周景亮; 曾绍锋; 易思敏
福建理工大学机械与汽车工程学院
Author(s):
YU Mingyuan;  ZHOU Jingliang;  ZENG Shaofeng;  YI Simin
School of Mechanical and Automotive Engineering, Fujian University of Technology
关键词:
轴承表面缺陷检测; YOLOv8; 注意力机制; 深度学习; 缺陷检测
Keywords:
bearing surface defect detection;  YOLOv8;  attention mechanisms;  deep learning;  defect detection
分类号:
TP391.4
DOI:
10.3969/j.issn.2097-3853.2024.03.011
文献标志码:
A
摘要:
针对深度学习模型在轴承表面缺陷检测过程中漏检率高、模型复杂度高的问题,提出了一种基于YOLOv8 的缺陷检测改进算法。在主干网络中引入GSConv 轻量化卷积模块,用GSConv 模块代替普通卷积,在不影响模型精度的情况下,减少模型的计算量;引入CBAM 卷积注意力模块,通过改进网络特征提取技术,提高了检测的准确性。实验结果表明,改进的模型在自建轴承表面缺陷检测数据集上的准确率为92.6%,较原模型准确率88.8%提高了3.8%;在提升准确率的同时,计算量也从8.2GFLOPs 减少到8.0 GFLOPs,证明了改进后的模型对轴承缺陷检测的有效性。
Abstract:
Aiming at the problems of high missing rate and high model complexity of deep learning model in the process of bearing surface defect detection, an improved defect detection algorithm based on YOLOv8 was proposed. Firstly, the GSConv lightweight convolutional module is introduced into the backbone network, and the GSConv module is used to replace the ordinary convolutional module, which reduces the calculation amount of the model without affecting the accuracy of the model. Secondly, CBAM convolutional attention module is introduced to improve the detection accuracy by improving the network feature extraction technology. Experimental results show that the accuracy of the improved model on the self-built bearing surface defect detection dataset is 92.6%, which is 3.8% higher than that of the original model (88.8%). While the accuracy is improved, the computational cost is also reduced from 8.2GFLOPs to 8.0 GFLOPs, which proves the effectiveness of the improved model for the detection of bearing defects.

参考文献/References:

［1］ 袁野，张群. 轴承套圈裂纹缺陷分析［J］. 金属世界，2021（3）：49-51. ［2］ 程锦锋，方贵盛，高惠芳. 表面缺陷检测的机器视觉技术研究进展［J］. 计算机应用研究，2023，40（4）：967-977. ［3］ 陈金贵，陈昊，张奔. 基于改进Niblack算法的轴承滚子表面缺陷检测［J］. 组合机床与自动化加工技术，2018（12）：82-85，97. ［4］ 陈硕，林志敏，吴岳彬，等. 轴承套圈端面缺陷在线视觉检测的研究与实现［J］. 轴承，2022（2）：48-54. ［5］ 罗会兰，陈鸿坤. 基于深度学习的目标检测研究综述［J］. 电子学报，2020，48（6）：1230-1239. ［6］ LIU Y，ZHANG C S，DONG X J. A survey of real-time surface defect inspection methods based on deep learning［J］. Artificial Intelligence Review，2023，56（10）：12131-12170. ［7］ 袁天乐，袁巨龙，朱勇建，等. 基于改进YOLOv5的推力球轴承表面缺陷检测算法［J］. 浙江大学学报（工学版），2022，56（12）：2349-2357. ［8］ 马忠平. 一种基于机器视觉的轴承环形表面缺陷检测［J］. 中国科技信息，2023（19）：129-132. ［9］ 苗荣慧,李志伟,武锦龙.基于改进YOLO v7的轻量化樱桃番茄成熟度检测方法［J］.农业机械学报,2023,54(10):225-233.［10］ 赵佰亭,吴俊东,贾晓芬.融合特征增强的轻量化罐道缺陷检测算法［J］.电子测量与仪器学报,2023,37(6):159-168. ［11］ WOO S，PARK J，LEE J Y，et al. CBAM：convolutional block attention module［C］∥European Conference on Computer Vision. Cham：Springer，2018：3-19.

更新日期/Last Update: 2024-06-25