Usage of an improved YOLOv5 for steel surface defect detection

Huihui Wen; Ying Li; Yu Wang; Haoyang Wang; Haolin Li; Hongye Zhang; Zhanwei Liu

doi:10.1515/mt-2023-0161

Article

Usage of an improved YOLOv5 for steel surface defect detection

Huihui Wen
Huihui Wen, lecturer, master tutor, Doctor of Mechanics, Beijing Institute of Technology. His main research areas are stress and image processing.
, Ying Li
Ying Li is a postgraduate student at Hebei University of Science and Technology. Her research interests include image processing and deep learning.
, Yu Wang
Yu Wang, a lecturer at Hebei University of Science and Technology, holds a master’s degree from Yanshan University. His main research interests are image processing and process control.
, Haoyang Wang
Haoyang Wang is a master’s candidate at Hebei University of Science and Technology, specializes in image stitching.
, Haolin Li
Haolin Li is a master’s candidate at Hebei University of Science and Technology. His research interests include target tracking and deep learning.
, Hongye Zhang
Hongye Zhang, who graduated from the Beijing Institute of Technology with a Ph. D. degree, is currently a master’s supervisor at Beijing Forestry University, specializing in resilience.
and Zhanwei Liu
Zhanwei Liu is a doctoral supervisor at the Beijing Institute of Technology. His main research interests are the study of resilience.

Published/Copyright: March 5, 2024

Published by

Become an author with De Gruyter Brill

Submit Manuscript Author Information

From the journal Materials Testing Volume 66 Issue 5

Abstract

The one-stage YOLOv5 steel surface defect detection has issues such as slow operation speed, loss of defect location and semantic information of small targets, and inadequate extraction of defect features. This study proposed a defect detection algorithm with improved YOLOv5 to solve these issues. The proposed algorithm used the slim-neck layer built by three new modules instead of the neck layer in YOLOv5s to achieve a lightweight network model. In addition, the spatial perception self-attention mechanism was introduced to enhance the feature extraction capability of the initial convolutional layer without limiting the input size. The improved Atrous Spatial Pyramid Pooling was added to expand the perceptual field and capture multiscale contextual information while preventing local information loss and enhancing the relevance of long-range information. The experimental results showed that the improved YOLOv5 algorithm has a reduced model volume, significantly higher detection accuracy and speed than the traditional algorithm, and the ability to detect steel surface defects quickly and accurately.

Keywords: steel; defect detection; atrous spatial pyramid pooling; spatial perception self-attention mechanism; slim-neck

Corresponding author: Yu Wang, School of Electrical Engineering, Hebei University of Science and Technology, Shijiazhuang 050000, Hebei, China, E-mail: wangyu0228@live.cn

Funding source: National Science and Technology Major Project

Award Identifier / Grant number: 2017-V1-0003-0073

Funding source: National Natural Science Foundation of China

Award Identifier / Grant number: 12002053,12002116,11972084

Award Identifier / Grant number: 12002116,12002053,11972084

About the authors

Huihui Wen

Huihui Wen, lecturer, master tutor, Doctor of Mechanics, Beijing Institute of Technology. His main research areas are stress and image processing.

Ying Li

Ying Li is a postgraduate student at Hebei University of Science and Technology. Her research interests include image processing and deep learning.

Yu Wang

Yu Wang, a lecturer at Hebei University of Science and Technology, holds a master’s degree from Yanshan University. His main research interests are image processing and process control.

Haoyang Wang

Haoyang Wang is a master’s candidate at Hebei University of Science and Technology, specializes in image stitching.

Haolin Li

Haolin Li is a master’s candidate at Hebei University of Science and Technology. His research interests include target tracking and deep learning.

Hongye Zhang

Hongye Zhang, who graduated from the Beijing Institute of Technology with a Ph. D. degree, is currently a master’s supervisor at Beijing Forestry University, specializing in resilience.

Zhanwei Liu

Zhanwei Liu is a doctoral supervisor at the Beijing Institute of Technology. His main research interests are the study of resilience.

Acknowledgments

Thanks to the above foundations for supporting this paper.

Research ethics: Not applicable.
Author contributions: Huihui Wen is a lecturer, master tutor, Doctor of Mechanics at Beijing Institute of Technology. His main research areas are stress and image processing. She contributed significantly to the conception of the paper. Ying Li is a postgraduate student at Hebei University of Science and Technology. She is the lead author of the paper. Yu Wang is a lecturer at Hebei University of Science and Technology, holds a master’s degree from Yanshan University. She mainly performs the acquisition, analysis or interpretation of work data. Haoyang Wang is a master’s candidate at Hebei University of Science and Technology, specializes in image stitching. His contribution is the critical revision of important intellectual content. Haolin Li is a master’s candidate at Hebei University of Science and Technology. His contribution was to draft the work. Hongye Zhang graduated from the Beijing Institute of Technology with a Ph. D. degree. His contribution was to finalize the paper. Zhanwei Liu is a doctoral supervisor at the Beijing Institute of Technology. His contribution is to ensure that issues relating to the accuracy or completeness of any part of the work are properly investigated and resolved.
Competing interests: The authors declare no conflicts of interest.
Research funding: Thanks for the support of the following fundings: National Natural Science Foundation of China (12002116, 12002053,11972084); National Natural Science Foundation of China (12002053,12002116,11972048); National Science and Technology Major Project (2017-V1-0003-0073). The funding organizations played no role in the study design, in the collection, analysis, and interpretation of data, in the writing of the report, or in the decision to submit the report for publication.
Data availability: Not applicable.

References

[1] C. Wu, Y. Z. Yu, L. H. Tam, and L. He, “Effects of bondline defects on the bond behaviour of CFRP-steel double strap joints,” Compos. Struct., vol. 308, 2023, Art. no. 116682, https://doi.org/10.1016/j.compstruct.2023.116682.Search in Google Scholar

[2] L. Sachin, D. Ravindra, S. Tushar, I. Mikhail, and B. Visvalingam, “Effect of deep cryogenic processing cycles on surface roughness, dimensional stability and microstructure of high carbon high chromium tool steel for cutting tool and dies applications,” Mater. Test., vol. 65, no. 4, pp. 629–640, 2022. https://doi.org/10.1515/mt.2022.0435.Search in Google Scholar

[3] J. Schwarzmann and T. Beiküfner, “Online magnetic flux leakage detection of inclusions and inhomogeneities in cold rolled steel plate,” Mater. Test., vol. 64, no. 10, pp. 1512–1526, 2022, https://doi.org/10.1515/mt.20220182.Search in Google Scholar

[4] C. Xu, C. Wu, L. Gao, Y. Xu, and G. Chen, “Detection of internal defects in CFRP strengthened steel structures using eddy current pulsed thermography,” Constr. Build. Mater., vol. 282, 2021, Art. no. 122642, https://doi.org/10.1016/j.conpstruct.2023.16682.Search in Google Scholar

[5] M. A. Machado, K. N. Antin, L. S. Rosado, P. Vilaca, and T. G. Santos, “High-speed inspection of delamination defects in unidirectional CFRP by non-contact eddy current testing,” Compos. B Eng., vol. 224, 2021, Art. no.109167, https://doi.org/10.1016/j.compositesb.2021.109167.Search in Google Scholar

[6] C. Feng, J. Kuo, S. H. Chen, and C. Y. Huang, “Automatic detection, classification and localization of defects in large photovoltaic plants using unmanned aerial vehicles (UAV) based infrared (IR) and RGB imaging,” Energy Convers. Manage., vol. 276, 2023, Art. no. 116495, https://doi.org/10.1016/j.enconman.2022.116495.Search in Google Scholar

[7] M. Puliti, G. Montaggioli, and A. Sabato, “Automated subsurface defects’ detection using point cloud reconstruction from infrared images,” Autom. Constr., vol. 129, no. 4, 2021, https://doi.org/10.1016/j.autcon.2021.103892.Search in Google Scholar

[8] Y. AboueNour and N. Gupta, “Assisted defect detection by in-process monitoring of additive manufacturing using optical imaging and infrared thermography,” Addit. Manuf., vol. 67, 2023, Art. no. 103483, https://doi.org/10.1016/j.addma.2023.1003483.Search in Google Scholar

[9] G. Dong, S. Sun, Z. X. Wang, N. S. Huang, P. N. Feng, and M. Q. Pan, “Application of machine vision-based NDT technology in ceramic surface defect detection – a review,” Mater. Test., vol. 64, no. 2, pp. 202–219, 2022, https://doi.org/10.1515/mt-2021-2012.Search in Google Scholar

[10] F. Liu, Y. Lei, X. R. Li, Q. M. Nan, Y. Lina, and L. Yue, “Vehicle identification using deep learning for expressway monitoring based on ultra-weak FBG arrays,” Opt. Express, vol. 31, no. 10, pp. 16754–16769, 2023, https://doi.1364/OE.487400.10.1364/OE.487400Search in Google Scholar PubMed

[11] Q. B. Zhou, R. Chen, B. Huang, W. Xu, J. Yu. “DeepInspection: Deep learning based hierarchical network for specular surface inspection,” Measurement vol. 160, 2020, Art. no. 107834, https://doi.org/10.1016/j.measurement.2020.107834.Search in Google Scholar

[12] F. Y. Zeng, et al.., “Rapid detection of white blood cells using hyperspectral microscopic imaging system combined with Multi-data Faster RCNN,” Sens. Actuators B, vol. 389, 2023, Art. no. 133865, https://doi.org/10.1016/j.snb.2023.133865.Search in Google Scholar

[13] M. Q. Chen, et al.., “Improved faster R-CNN for fabric defect detection based on Gabor filter with Genetic Algorithm optimization,” Comput. Ind., vol. 134, 2022, Art. no. 103551, https://doi.org/10.1016/j.compind.2021.103551.Search in Google Scholar

[14] K. B. Chen, Y. Xuan, A. J. Lin, and S. H. Guo, “Esophageal cancer detection based on classification of gastrointestinal CT images using improved Faster RCNN,” Comput. Methods Programs Biomed., vol. 207, 2021 , Art no. 106172, https://doi.org/10.1016/j.cmpb.2021.106172.Search in Google Scholar PubMed

[15] D. D. Wang and D. G. He, “Fusion of mask RCNN and attention mechanism for instance segmentation of apples under complex background,” Comput. Electron. Agric., vol. 196, 2022, Art. no. 106864, https://doi.org/10.1016/.compag.2022.106864.Search in Google Scholar

[16] B. Su, H. Chen, and Z. Zhou, “BAF-detector: an efficient CNN-based detector for photovoltaic cell defect detection,” IEEE Trans. Ind. Electron., vol. 99, 2021.10.1109/TIE.2021.3070507Search in Google Scholar

[17] C. Qi, J. F. Gao, S. Person, H. Harman, K. Chen, and L. Shu, “Tea chrysanthemum detection under unstructured environments using the TC-YOLO mode,” Expert Syst. Appl., vol. 193, no. 4, pp. 121–130, 2022, https://doi.org/10.1016/j.eswa.2021.116473.Search in Google Scholar

[18] Q. W. Qiu and D. Lau, “Real-time detection of cracks in tiled sidewalks using YOLO-based method applied to unmanned aerial vehicle (UAV) images,” Autom. Constr., vol. 147, 2023, Art. no. 104745, https://doi.org/10.1016/j.autcon.2023.104745.Search in Google Scholar

[19] Z. Huang, J. L. Wang, X. S. Fu, T. Yu, Y. Q. Guo, and R. U. Wang, “DC-SPP-YOLO: dense connection and spatial pyramid pooling based YOLO for object detection,” Inf. Sci., vol. 522, pp. 241–258, 2020, https://doi.org/10.1016/j.ins.2020.02.067.Search in Google Scholar

[20] Y. Tan, R. Y. Cai, P. L. Chen, and M. Z. Wang, “Automatic detection of sewer defects based on improved you only look once algorithm,” Autom. Constr., vol. 131, 2021, Art. no. 103912, https://doi.org/10.1016/j.autcon.2021.103912.Search in Google Scholar

[21] G. Yang, C. H. Song, Z. J. Yang, and S. P. Cui, “Bubble detection in photoresist with small samples based on GAN augmentations and modified YOLO,” Eng. Appl. Artif. Intell., vol. 123, p. 106224, 2023, Art. no. 106224, https://doi.org/10.1016/j.engappai.2023.Search in Google Scholar

Published Online: 2024-03-05

Published in Print: 2024-05-27

You are currently not able to access this content.

Articles in the same Issue

https://doi.org/10.1515/mt-2023-0161

Keywords for this article

steel; defect detection; atrous spatial pyramid pooling; spatial perception self-attention mechanism; slim-neck