摘要
针对传统钢板表面缺陷检测方法效果差、缺陷定位不准确等问题,提出一种基于改进RetinaNet-GHM的深度学习检测算法。首先,引入路径聚合特征金字塔网络融合浅层和深层语义信息,提升网络对小目标的检测效果;然后,使用GHMC和GHMR损失函数对缺陷进行分类和定位;最后,引入高斯形式的软化非极大值抑制算法,提高检测精度。实验结果表明,改进的RetinaNet-GHM算法的平均精度均值为76.7%,裂纹、夹杂、斑块、麻点、压入氧化铁皮以及划痕六类缺陷的平均精度分别为45.2%、88.2%、94.2%、86.1%、65.1%和87.4%。通过与其他经典算法相比,改进的RetinaNet-GHM算法具有较好的检测效果.
Aiming at the problems of poor effect and inaccurate defect location of traditional steel plate surface defect detection methods,a deep learning detection algorithm based on improved RetinaNet-GHM was proposed.Firstly,the path aggregation feature pyramid network is introduced to fuse shallow and deep semantic information to improve the detection effect of the network on small targets.Then,GHMC and GHMR loss functions are used to classify and locate defects.Finally,the soft-non maximum suppression algorithm in Gaussian form is introduced to improve the detection accuracy.The experimental results show that the average accuracy of the improved RetinaNet-GHM algorithm is 76.7%,and the average accuracy of crazing,inclusion,patches,pitted surface,rolled-in_scale and scratchs is 45.2%,88.2%,94.2%,86.1%,65.1%and 87.4%respectively.Compared with other classical algorithms,the improved RetinaNet-GHM algorithm has better detection effect.
作者
李雪露
杨永辉
储茂祥
Li Xuelu;Yang Yonghui;Chu Maoxiang(School of Electronic and Information Engineering,University of Science and Technology Liaoning,Anshan 114051,China)
出处
《电子测量技术》
北大核心
2023年第6期100-105,共6页
Electronic Measurement Technology
基金
国家自然科学基金(21978123)
辽宁省高等学校基本科研项目(2020LNZD06)资助
作者简介
李雪露,硕士研究生,主要研究方向为图像处理、模式识别。E-mail:lxl382636745@163.com;通信作者:杨永辉,教授,博士,博士生导师,主要研究方向为模式识别与智能控制。E-mail:yangyh2636688@163.com;储茂祥,教授,博士,博士生导师,主要研究方向为模式识别与智能控制。E-mail:chu52_2004@163.com
