Surface Damage Detection Method for Retired Shaft Parts Based on Improved YOLOv5

doi:10.12068/j.issn.1005-3026.2024.07.012

Abstract

Abstract:

Aiming at the existing problems of low efficiency and poor consistency in the damage detection of retired shaft parts by traditional detection methods， an improved YOLOv5?based surface damage detection method for retired shaft parts is proposed. Firstly， the attention mechanism is embedded into the detection algorithm to enhance the feature representation of damage in the image. Then the network structure of the detection model is improved by using the repeated weighted bidirectional feature fusion method to effectively enhance the network feature extraction capability. Finally， Ghostconv convolution module is used instead of normal convolution， which drastically reduces the number of model parameters. The experimental results show that the accuracy of the modified algorithm model has improved by 6.9% compared to the original YOLOv5， reaching 88.4%， while the number of model parameters has reduced by 6.1%， ensuring the detection speed is on par with YOLOv5. Compared with such mainstream detection methods as YOLOv3， SSD and Faster-RCNN， its detection accuracy has a significant advantage while ensuring a higher detection speed.

Key words: YOLOv5, surface damage detection, attention mechanism, multiplex feature fusion, Ghostconv

CLC Number:

TP 391.4

Wei-wei LIU, Jia-he QIU, Guang-da HU, Ze-yuan LIU. Surface Damage Detection Method for Retired Shaft Parts Based on Improved YOLOv5[J]. Journal of Northeastern University(Natural Science), 2024, 45(7): 1002-1010.

Figures/Tables 13

Fig.1 YOLOv5 network structure

Fig.2 Structure of CA module

Fig.3 Network architecture

Fig.4 Structure of the improved YOLOv5

Fig.5 Model training results

Fig.6 Type of damage

Fig.7 Statistics on the number of samples for each defect in the data set

Table 1 Ablation experiments of YOLOv5

算法	CA	BiFPN	Ghostconv	精度	召回率	mAP/%	参数量	F1	检测速度/(帧 $·$ s^-1)
YOLOv5				0.840	0.91	80.0	7 030 417	0.76	31
Improved 1	√			0.885	0.94	84.7	7 045 521	0.82	29
Improved 2		√		0.907	0.93	84.9	7 169 402	0.84	28
Improved 3	√	√		0.917	0.96	86.6	7 112 097	0.83	24
Improved 4		√	√	0.875	0.96	86.0	6 665 946	0.83	30
Improved 5	√	√	√	0.873	0.98	88.4	6 679 601	0.84	33

Table 1 Ablation experiments of YOLOv5

算法	CA	BiFPN	Ghostconv	精度	召回率	mAP/%	参数量	F1	检测速度/(帧 $·$ s^-1)
YOLOv5				0.840	0.91	80.0	7 030 417	0.76	31
Improved 1	√			0.885	0.94	84.7	7 045 521	0.82	29
Improved 2		√		0.907	0.93	84.9	7 169 402	0.84	28
Improved 3	√	√		0.917	0.96	86.6	7 112 097	0.83	24
Improved 4		√	√	0.875	0.96	86.0	6 665 946	0.83	30
Improved 5	√	√	√	0.873	0.98	88.4	6 679 601	0.84	33

Fig.8 P-R curves of each method in the ablation experiment

Fig.9 Comparison of detection results before and after improvement

Table 2 Comparison of the attention mechanism module performance

算法	SE	CBAM	CA	精度	召回率	mAP/%	参数量	F1	检测速度/(帧·s^-1)
Model 1	√			0.884	0.97	85.9	6 671 377	0.84	32
Model 2		√		0.907	0.95	86.8	6 671 475	0.85	33
本文			√	0.873	0.98	88.4	6 679 601	0.84	33

Fig.10 P-R curves of three detection methods

Table 3 Performance comparison of different models

算法	速度/(帧·s^-1)	mAP/%	AP/%
算法	速度/(帧·s^-1)	mAP/%	凹坑	变形	划痕	锈蚀
YOLOv3	29	69.7	83.3	69.9	64.6	51.4
SSD	18	76.2	76.2	93.4	73.1	62.4
Faster-RCNN	<10	64.6	77.6	67.3	59.3	54.5
Faster-RCNN(FPN)	<10	80.3	87.4	85.5	83.9	64.4
YOLOv5	32	80.0	90.5	97.0	81.3	51.1
Improved YOLOv5	33	88.4	98.4	90.3	87.3	77.7

References 17

1	Syed R， Wang J G， Jing D Y，et al.Case study：optimization of case depth in induction‑hardened 42CrMo steel shaft［J］.IOP Conference Series：Materials Science and Engineering，2020，831（1）：012004.
2	孙斌.金属轴类零件表面缺陷成像与判识技术研究［D］.南京：南京理工大学，2017.
	Sun Bin.Research on surface defect imaging and identification technology of metal shaft parts［D］.Nanjing：Nanjing University of Science and Technology，2017.
3	Jeon I， Lim H J， Liu P P，et al.Fatigue crack detection in rotating steel shafts using noncontact ultrasonic modulation measurements［J］.Engineering Structures，2019，196：109293.
4	Neslušan M， Bahleda F， Minárik P，et al.Non‑destructive monitoring of corrosion extent in steel rope wires via Barkhausen noise emission［J］.Journal of Magnetism and Magnetic Materials，2019，484：179-187.
5	Sha J W， Fan M B， Cao B H，et al.Noncontact and nondestructive evaluation of heat‑treated bearing rings using pulsed eddy current testing［J］.Journal of Magnetism and Magnetic Materials，2021，521：167516.
6	Xie S J， Zhang L， Zhao Y，et al.Features extraction and discussion in a novel frequency‑band‑selecting pulsed eddy current testing method for the detection of a certain depth range of defects［J］.NDT & E International，2020，111：102211.
7	Dai J J， Li T P， Xuan Z L，et al.Automated defect analysis system for industrial computerized tomography images of solid rocket motor grains based on YOLO‑V4 model［J］.Electronics，2022，11（19）：3215.
8	Tang M， Li Y Y， Yao W，et al.A strip steel surface defect detection method based on attention mechanism and multi‑scale maxpooling［J］.Measurement Science and Technology，2021，32（11）：115401.
9	Huang Z Y， Hu H J， Shen Z Y，et al.Lightweight edge‑attention network for surface‑defect detection of rubber seal rings［J］.Measurement Science and Technology，2022，33（8）：085401.
10	Guo S Y， Li L L， Guo T Y，et al.Research on mask‑wearing detection algorithm based on improved YOLOv5［J］.Sensors，2022，22（13）：4933.
11	李鑫，李香蓉，汪诚，等.基于改进YOLOv5的航空发动机表面缺陷检测模型［J］.激光与光电子学进展，2023，60（16）：1615007.
	Li Xin， Li Xiang‑rong， Wang Cheng，et al.Aero‑engine surface defect detection model based on improved YOLOv5［J］.Laser & Optoelectronics Progress，2023，60（16）：1615007.
12	Hou Q B， Zhou D Q， Feng J S.Coordinate attention for efficient mobile network design［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Nashville，2021：13713-13722.
13	Tan M X， Pang R M， Le Q V.EfficientDet：scalable and efficient object detection［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle，2020：10781-10790.
14	Han K， Wang Y H， Tian Q，et al.Ghostnet：more features from cheap operations［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle，2020：1580-1589.
15	Liu W， Anguelov D， Erhan D，et al.SSD：single shot multibox detector［C］// Proceedings of 14th European Conference on Computer Vision-ECCV 2016.Cham：Springer，2016：21-37.
16	Redmon J， Farhadi A.YOLOv3：an incremental improvement［EB/OL］.（2018-04-08）［2023-03-01］..
17	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，2017，39（6）：1137-1149.

[1]	An-lin TIAN, Wei-min LEI, Peng ZHANG, Wei ZHANG. A Multi-scale Edge Detection Method Based on Encoder-Decoder [J]. Journal of Northeastern University(Natural Science), 2024, 45(7): 936-943.
[2]	Li-xin GUO, Su-tao BI, Ming-yang ZHAO. State Detection Algorithm of Manipulator Based on Improved YOLOv4 Lightweight Network [J]. Journal of Northeastern University(Natural Science), 2024, 45(6): 769-775.
[3]	Yuan MA, Li-huang SHE, Jia-wei LI, Xi-rong BAO. Adaptive Graph Convolutional 3D Point Cloud Recognition Algorithm Based on Attention Mechanism [J]. Journal of Northeastern University(Natural Science), 2024, 45(6): 786-792.
[4]	Hu FENG, Ke-chen SONG, Wen-qi CUI, Yun-hui YAN. Few-Shot Semantic Segmentation of Strip Steel Surface Defects Based on Meta-Learning [J]. Journal of Northeastern University(Natural Science), 2024, 45(3): 354-360.
[5]	Ying SUN, Ya-ru ZHOU, Xue-ying ZHANG. Speech Emotion Recognition Fusing Functional Paralanguage Proportion Coefficient [J]. Journal of Northeastern University(Natural Science), 2024, 45(1): 40-48.
[6]	JIANG Yang， LIU Cheng， DING Qi-chuan， WANG Li. Segmentation of COVID-19 CT Images Based on Dual Attention Mechanism [J]. Journal of Northeastern University(Natural Science), 2023, 44(9): 1259-1268.
[7]	ZHOU Song， GAO Tian-han. EEG Recognition Method for Epileptic Patients Based on RNN Model with Attention Mechanism [J]. Journal of Northeastern University(Natural Science), 2023, 44(8): 1098-1103.
[8]	HAO Bo， YIN Xing-chao， YAN Jun-wei， ZHANG Li. Gesture Recognition in the Complex Environment Based on Gan-St-YOLOv5 [J]. Journal of Northeastern University(Natural Science), 2023, 44(7): 953-963.
[9]	DING Qi-chuan， WANG Li， LIU Cheng. Classification of Pulmonary Nodule by Combining Long-Distance Channel Attention and Pathological Feature [J]. Journal of Northeastern University(Natural Science), 2023, 44(4): 476-485.
[10]	CHEN Cheng， SHI Pei-xin， WANG Zhan-sheng， JIA Peng-jiao. Shield Load Prediction Method Based on Deep Learning with Multiattention Mechanism [J]. Journal of Northeastern University(Natural Science), 2023, 44(11): 1631-1638.
[11]	LIN Qing-yang， CHEN Xiao-fang， XIE Yong-fang. An Superheat Identification Method in Aluminium Electrolysis Based on Residual Convolutional Self-Attention Neural Network [J]. Journal of Northeastern University(Natural Science), 2023, 44(1): 8-17.
[12]	LIU Yang， YAN Dong-mei， MENG Fan-wei. Improved Two-Branch Person Re-identification Algorithm Based on Transformer [J]. Journal of Northeastern University(Natural Science), 2023, 44(1): 26-32.
[13]	GU De-ying， LUO Yu-lun， LI Wen-chao. Traffic Target Detection in Complex Scenes Based on Improved YOLOv5 Algorithm [J]. Journal of Northeastern University(Natural Science), 2022, 43(8): 1073-1079.
[14]	DAI Yin， LIU Wei-bin， DONG Xin-yang， SONG Yu-meng. U-Net CSF Cells Segmentation Based on Attention Mechanism [J]. Journal of Northeastern University(Natural Science), 2022, 43(7): 944-950.
[15]	YU Zhe-zhou， LIU Yan， LIU Yuan-ning. Improved Iris Locating Algorithm Based on YOLOV3 [J]. Journal of Northeastern University(Natural Science), 2022, 43(4): 496-501.