Improved Steel Defect Detection Method Based on Enhanced Fusion of RFB and YOLOv5 Features

doi:10.19678/j.issn.1000-3428.0068476

Abstract

Abstract:

Factors such as irregular shapes, different sizes, and complex backgrounds of steel surface defects significantly increase the difficulty in detecting steel surface defects. To overcome the drawbacks of existing methods, such as low detection accuracy, low detection speed, and difficulty in detecting small target defects, an improved steel defect detection method, namely RFB-YOLOv5-E, based on the enhanced fusion of Receptive Field Block (RFB) and YOLOv5 features, is proposed to improve the recognition rate of steel surface defects. First, the C3 module in YOLOv5 is modified and upgraded to the C3s module to obtain more gradient information by adding more gradient flow branches, which improves the accuracy of the model. Second, the shallow feature extraction network is improved, and feature enhancement functions are added, which increase the difference between background and defect. A downsampling layer and a detection head are added to expand the depth and receptive field of the network in order to improve the feature extraction and detection capabilities. In addition, the RFB module is improved, and it replaces the Spatial Pyramid Pooling-Fast (SPPF) module in the YOLOv5 backbone network. By simulating human vision, the receptive field is further enlarged, and the feature extraction capability of the network is further enhanced. The experimental results show that the mean Average Precision (A_mAP) of the RFB-YOLOv5-E algorithm on the NEU-DET dataset reaches 79.2%, which is 8.5% higher than that of the original YOLOv5 algorithm. Further, the detection speed is 122 frames per second, which indicates that a better balance between detection speed and detection accuracy is achieved.

Key words: defect detection, feature enhancement, receptive field block, YOLOv5 method

摘要：

钢材表面缺陷形状不规则、尺度不一、背景复杂等因素大大增加了钢材表面缺陷检测的难度。针对现有方法检测精度低、检测速度小、小目标缺陷难以检测等问题, 提出了一种基于感受野块(RFB)和YOLOv5特征增强融合改进的钢材表面缺陷检测(RFB-YOLOv5-E)模型, 以提高对钢材表面缺陷的识别率。首先, 对YOLOv5中的C3模块进行修改, 将其升级为C3s模块, 通过增加更多的梯度流分支来获取更多的梯度信息, 从而提高模型的准确度; 然后改进浅层特征提取网络, 添加特征增强函数以增大背景与缺陷间的差距, 再增加一个下采样层和一个检测头以扩大网络的深度和感受野, 进而提高特征提取的能力和检测能力; 此外, 还改进了RFB并替换YOLOv5主干网络中的空间金字塔池化(SPPF)模块, 通过模拟人类视觉进一步增大感受野, 进一步强化网络的特征提取能力。实验结果表明, RFB-YOLOv5-E算法在NEU-DET数据集上的均值平均精度(A_mAP)达到了79.2%, 较原YOLOv5算法提高了8.5%, 检测速度为122帧/s, 实现了检测速度与检测精度更好的均衡。

关键词: 缺陷检测, 特征增强, 感受野块, YOLOv5方法

HUANG Shuoqing, HUANG Jingui. Improved Steel Defect Detection Method Based on Enhanced Fusion of RFB and YOLOv5 Features[J]. Computer Engineering, 2025, 51(4): 249-260.

黄硕清, 黄金贵. 基于RFB和YOLOv5特征增强融合改进的钢材表面缺陷检测方法[J]. 计算机工程, 2025, 51(4): 249-260.

/ Recommend / Download Citations

URL: https://www.ecice06.com/EN/10.19678/j.issn.1000-3428.0068476

https://www.ecice06.com/EN/Y2025/V51/I4/249

Figures/Tables 17

Fig.1 YOLOv5 network structure diagram

Fig.2 Dilated convolution

Fig.3 RFB-YOLOv5-E model structure

Fig.4 C3 module is improved to C3s module

Fig.5 CBS module with feature enhancement function

Fig.6 Comparison of final detection results and shallow feature maps

Fig.7 Improved RFB structure

Fig.8 A_mAP of Soft-NMS at different thresholds

Fig.9 Graphs of test results of various categories

References 25

1	KIM S, KIM W, NOH Y K, et al. Transfer learning for automated optical inspection[C]//Proceedings of 2017 International Joint Conference on Neural Networks. Washington D. C., USA: IEEE Press, 2017: 2517-2524.
2	GIRSHICK R. Fast R-CNN[C]// Proceedings of International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2015: 1440-1448.
3	REN S , HE K , GIRSHICK R , et al. Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 39 (6): 1137- 1149.
4	REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2016: 779-788.
5	REDMON J, FARHADI A. YOLO9000: Better, Faster, Stronger[C]//Proceedings of IEEE Conference on Computer Vision & Pattern Recognition. Washington D. C., USA: IEEE Press, 2017: 6517-6525.
6	REDMON J, FARHADI A. YOLOv3: an incremental improvement[EB/OL]. (2018-04-08)[2023-07-27]. https://arxiv.org/abs/1804.02767.
7	BOCHKOVSKIY A, WANG C Y, LIAO H. YOLOv4: optimal speed and accuracy of object detection[EB/OL]. (2018-04-08)[2023-07-27]. https://arxiv.org/abs/2004.10934.
8	LIN T Y, DOLLÁR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2017: 2117-2125.
9	LIU S, QI L, QIN H, et al. Path aggregation network for instance segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 8759-8768.
10	REZATOFIGHI H, TSOI N, GWAK J Y, et al. Generalized intersection over union: a metric and a loss for bounding box regression[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2019: 658-666.
11	KOU X , LIU S , CHENG K , et al. Development of a YOLO-V3-based model for detecting defects on steel strip surface. Measurement, 2021, 182, 109454. doi: 10.1016/j.measurement.2021.109454
12	曹义亲, 伍铭林, 徐露. 基于改进YOLOv5算法的钢材表面缺陷检测. 图学学报, 2023, 44 (2): 335- 345.
	CAO Y Q , WU M L , XU L . Steel surface defect detection based on improved YOLOv5 algorithm. Journal of Graphics, 2023, 44 (2): 335- 345.
13	王浩然. 基于YOLOv5的钢材表面缺陷检测研究[D]. 桂林: 广西师范大学, 2022.
	WANG H R. Research on steel surface defects detection based on YOLOv5[D]. Guilin: Guangxi Normal University, 2022.
14	ZHENG Z, WANG P, LIU W, et al. Distance-IoU loss: faster and better learning for bounding box regression[C]//Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto, USA: AAAI, 2020: 12993-13000.
15	GE Z, LIU S, WANG F, et al. YOLOX: exceeding YOLO series in 2021[EB/OL]. (2021-08-06)[2023-07-27]. https://arxiv.org/abs/2107.08430?context=cs.C.
16	WANG C Y, BOCHKOVSKIY A, LIAO H. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]// Proceedings of Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2023: 1-10.
17	YU F, KOLTUN V. Multi-scale context aggregation by dilated convolutions[EB/OL]. (2016-04-30)[2023-07-27]. https://arxiv.org/abs/1511.07122.
18	NEUBECK A, VAN GOOL L. Efficient non-maximum suppression[C]//Proceedings of 18th International Conference on Pattern Recognition (ICPR'06). Washington D. C., USA: IEEE Press, 2006: 850-855.
19	BODLA N, SINGH B, CHELLAPPA R, et al. Soft-NMS-improving object detection with one line of code[C]//Proceedings of the IEEE International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2017: 5561-5569.
20	WANG C Y, LIAO H Y M, WU Y H, et al. CSPNet: a new backbone that can enhance learning capability of CNN[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Washington D. C., USA: IEEE Press, 2020: 390-391.
21	LIU S, HUANG D. Receptive field block Net for accurate and fast object detection[C]// Proceedings of the European Conference on Computer Vision (ECCV). Berlin, Germany: Springer, 2018: 385-400.
22	SZEGEDY C, IOFFE S, VANHOUCKE V, et al. Inception-v4, inception-ResNet and the impact of residual connections on learning[C]//Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto, USA: AAAI, 2017: 4278-4284.
23	KRIZHEVSKY A , SUTSKEVER I , HINTON G E . ImageNet classification with deep convolutional neural networks. Communications of the ACM, 2017, 60 (6): 84- 90. doi: 10.1145/3065386
24	SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[EB/OL]. (2015-04-10)[2023-07-27]. https://arxiv.org/abs/1409.1556.
25	HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2016: 770-778.

[1]	LI Dahai, LÜ Chungui, WANG Zhendong. Scene Text Image Super-Resolution Reconstruction Based on Dual-Branched Sequence Residual Attention [J]. Computer Engineering, 2024, 50(9): 286-295.
[2]	Yifeng WANG, Cheng ZENG, Qingyu QUAN, Jiaoran WANG, Peng HE. Smart Contract Defect Detection Method Based on Multi-Feature Fusion [J]. Computer Engineering, 2024, 50(8): 133-141.
[3]	Shijing ZHANG, Xutao MO, Xing ZHAO, Yanglin DONG. Internal Thread Defect Detection Based on Spherical Catadioptric Imaging and YOLOv7 [J]. Computer Engineering, 2024, 50(7): 282-292.
[4]	ZHANG Huiyan, LIANG Yong, LAN Jinghong, ZHAO Qiang. Intrusion Detection Method Based on Memory Module and Filtered Generative Adversarial Network [J]. Computer Engineering, 2024, 50(6): 197-207.
[5]	SUN Bangyong, MA Ming, YU Tao. Multi-scale Camouflaged Object Detection Method Based on Regional Feature Enhancement [J]. Computer Engineering, 2024, 50(5): 209-219.
[6]	CUI Liqun, CAO Huawei. Target Detection of Remote-Sensing Images Based on Improved YOLOv5 [J]. Computer Engineering, 2024, 50(4): 228-236.
[7]	BAI Yu, WANG Jun, RAN Honglei, AN Shengbiao. Research on Internal Defect Annotation and Detection Methods of Semiconductor Devices [J]. Computer Engineering, 2024, 50(12): 245-253.
[8]	ZHU Yanbin, WANG Runmin, CHEN Hua, CAO Xiaofei, ZHU Zhenlin, DING Yajun. Traffic Text Detection Method Based on Multi-Granularity Feature Enhancement Network [J]. Computer Engineering, 2024, 50(11): 80-88.
[9]	Ang LU, Jun CHU, Lu LENG. Image Dehazing Based on High-Frequency and Low-Frequency Feature Enhancement [J]. Computer Engineering, 2023, 49(8): 174-181.
[10]	Shan WU, Feng ZHOU. Small Target Detection Based on Improved SSD Algorithm [J]. Computer Engineering, 2023, 49(7): 179-188.
[11]	Juanjuan JI, Jia WANG, Yajie CHEN, Daohua LU. Detection of Minor Defects on the Surface of Hydraulic Valve Block Based on Improved YOLOv7 [J]. Computer Engineering, 2023, 49(11): 302-310.
[12]	Gang LI, Rui SHAO, Mingle ZHOU, Min LI, Honglin WAN. Lightweight Industrial Products Defect Detection Network Based on Attention [J]. Computer Engineering, 2023, 49(11): 275-283.
[13]	Jing ZHOU, Yuan ZHONG, Ping LI, Yi YANG, Litai MA, Tao ZHANG. Multi-Scale Feature Fusion Attention Network Model for Cervical Vertebrae MRI Segmentation [J]. Computer Engineering, 2023, 49(10): 298-304.
[14]	HU Xinrong, GONG Chuang, ZHANG Zili, ZHU Qiang, PENG Tao, HE Ruhan. Clothing Image Segmentation Network Based on Improved Deeplab v3+ [J]. Computer Engineering, 2022, 48(7): 284-291.
[15]	ZHANG Yexing, CHEN Min, PAN Qiuyu. Object Detection Method Based on Feature Channel Modeling [J]. Computer Engineering, 2022, 48(7): 264-269,299.

Please choose a citation manager

Content to export