您好,欢迎来到中国测试科技资讯平台!

首页> 《中国测试》期刊 >本期导读>基于改进YOLOv4的铝材表面缺陷检测方法

基于改进YOLOv4的铝材表面缺陷检测方法

457    2024-05-24

免费

全文售价

作者:李澄非, 蔡嘉伦, 邱世汉, 梁辉杰, 徐傲

作者单位:五邑大学智能制造学部,广东 江门 529020


关键词:目标检测;铝材表面缺陷;YOLOv4;注意力机制;机器视觉


摘要:

针对铝材表面缺陷检测精度不高,容易漏检的问题,提出基于改进YOLOv4的缺陷检测方法。在CSPResblock模块中引入注意力机制SE模块,赋予各个通道相应的权重,加强网络对于重要信息的训练,提升网络的特征提取能力;改进SPP模块,使用不同宽高比的池化核,有利于保留更多的短边信息,提高网络对大宽高比缺陷的检测能力;对PANet结构进行改进,在对应特征层级上拼接输入信息,主要融合主干网络的三层输出,获得更多较浅的特征信息,提升对小目标的检测能力;实验结果表明,改进后的YOLOv4算法在铝材表面缺陷数据集上的精度(mAP)达79.27%,优于其他常见目标检测算法。


Defect detection method in aluminum material surface based on improved YOLOv4
LI Chengfei, CAI Jialun, QIU Shihan, LIANG Huijie, XU Ao
Faculty of Intelligent Manufacturing, Wuyi University, Jiangmen 529020, China
Abstract: Aiming at the problem of low accuracy and easily missing defects in aluminum surface defect detection, a defect detection method of improved YOLOv4 is proposed. SE is introduced into CSPResblock giving channels weights, which can increase training effect for important information and improve the ability of feature extraction. Using pooling kernels with different aspect ratios is beneficial to retain more short-side information, so SPP is revised to improve the network's ability to detect large aspect ratio defects. PANet is improved to fuse more input shallow feature information from the three outputs of backbone, increasing the ability in detecting small objects. The experiment result shows that mAP of improved YOLOv4 algorithm achieves 79.27% in aluminum surface defect data set, better than other common object detection algorithms.
Keywords: object detection;aluminum material surface defect;YOLOv4;attention mechanism;machine vision
2024, 50(5):160-166  收稿日期: 2022-05-21;收到修改稿日期: 2022-07-01
基金项目: 广东省科技发展专项资金(2017A010101019);广东省普通高校特色创新类项目(2019KTSCX181);广东省研究生教育创新计划项目(2021JGXM109)
作者简介: 李澄非(1971-),女,河南南阳市人,副教授,博士,研究方向为图像信息处理、机器视觉、复杂工业过程故障诊断及检测。
参考文献
[1] DALAL N, TRIGGS B. Histograms of oriented gradients for human detection[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, 2005: 886-893.
[2] LOWE D G. Distinctive image features from scale invariant key points[J]. International Journal of Computer Vision, 2004, 60(2): 581-588
[3] VAPNIK V N. The nature of statistical learning theory[J]. Technometrics, 1997, 8(6): 1564
[4] SUVDAA B, AHN J, KO J. Steel surface defects detection and classification using SIFT and voting strategy[J]. International Journal of Software Engineering and Its Applications, 2012, 6(2): 161-166
[5] 郭慧, 徐威, 刘亚菲. 基于支持向量机的钢板表面缺陷检测[J]. 东华大学学报(自然科学版), 2018, 44(4): 635-639
GUO H, XU W, LIU Y F. Steel plate surface defect detection based on support vector machine[J]. Journal of Donghua University(Natural Science), 2018, 44(4): 635-639
[6] 殷鹰, 谢罗峰, 黄泰博. 基于深度学习的磁瓦内部缺陷声振检测方法[J]. 中国测试, 2020, 46(3): 32-38
YIN Y, XIE L F, HUANG T B. A deep learning method for magnetic tile internal defect inspection based on acoustic vibration[J]. China Measurement & Test, 2020, 46(3): 32-38
[7] 伊欣同, 单亚峰. 基于改进Faster R-CNN的光伏电池内部缺陷检测[J]. 电子测量与仪器学报, 2021, 35(1): 40-47
YI X T, DAN Y F. Internal defect detection of photovoltaic cells based on improved Faster R-CNN[J]. Journal of Electronic Measurement and Instrument, 2021, 35(1): 40-47
[8] 彭伟康, 陈爱军, 吴东明, 等. 基于改进Faster R-CNN的水准泡缺陷检测方法[J]. 中国测试, 2021, 47(7): 6-12
PENG W K, CHEN A J, WU D M, et al. Defect detection method of level bubble based on improved Faster R-CNN[J]. China Measurement & Test, 2021, 47(7): 6-12
[9] 程婧怡, 段先华, 朱伟. 改进YOLOv3的金属表面缺陷检测研究[J]. 计算机工程与应用, 2021, 57(19): 252-258
CHENG J Y, DUAN X H, ZHU W. Research on improving metal surface defect detection of YOLOv3[J]. Computer Engineering and Applications, 2021, 57(19): 252-258
[10] 吴越, 杨延竹, 苏雪龙, 等. 基于Faster R-CNN的钢板表面缺陷检测方法[J]. 东华大学学报(自然科学版), 2021, 47(3): 84-89
WU Y, YANG Y Z, SU X L, et al. Surface defect detection method for steel plates based on Faster R-CNN[J]. Journal of Donghua University(Natural Science), 2021, 47(3): 84-89
[11] 李庆党, 李铁林. 基于改进YOLOv3算法的钢板缺陷检测[J]. 电子测量技术, 2021, 44(2): 104-108
LI Q D, LI T L. Steel plate defect detection based on improved YOLOv3 algorithm[J]. Electronic Measurement Technology, 2021, 44(2): 104-108
[12] REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN: Towards real-time object detection with region proposal networks[J]. IEEE transactions on pattern analysis and machine intelligence, 2016, 39(6): 1137-1149
[13] BOCHKOVSKIY A, WANG C Y, LIAO H Y M. Yolov4: Optimal speed and accuracy of object detection[C]//IEEE conference on Computer Vision and Pattern Recognition, 2020.
[14] REDMON J, FARHADI A. Yolov3: An incremental improvement[C]//Proceedings of 2018 IEEE Conference on Computer Vision and Pattern Recognition, 2018: 1-6.
[15] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 7132-7141.