基于改进YOLOv7-tiny的橡胶密封圈缺陷检测方法

doi:10.11996/JG.j.2095-302X.2024030446

摘要/Abstract

摘要：

针对橡胶密封圈表面缺陷传统检测效率低下的问题，提出一种改进YOLOv7-tiny的橡胶密封圈表面缺陷检测算法。在主干特征提取网络中引入PConv优化ELAN结构，增强算法特征提取能力，并减少参数量；在特征融合网络中引入全局注意力机制(GAM)，利用每一对三维通道、空间宽度和空间高度之间的注意力权重，在3个维度上捕捉重要特征来提高效率，增强算法特征融合能力；使用WIoU损失函数优化原边界框损失函数，通过符合情况的梯度增益分配策略，增强算法对检测目标的定位能力；增加P2小目标检测层，加强深层与浅层特征信息的融合，增强算法对小目标缺陷的检测能力。在O-Rings数据集进行实验对比，改进后的算法与YOLOv7-tiny算法比较，mAP提升了7.8%，达到了90.9%的检测精度，能够满足实际工业生产需求。

关键词: YOLOv7-tiny, 橡胶密封圈, 缺陷检测, 注意力机制, 小目标检测层

Abstract:

Aiming at the problem of low efficiency in traditional detection of surface defects of rubber seal rings, an improved YOLOv7-tiny algorithm for surface defect detection of rubber seal rings was proposed. The PConv optimized ELAN structure was introduced into the backbone feature extraction network to enhance the algorithm’s feature extraction capability and to reduce the number of parameters. The global attention mechanism (GAM) was introduced into the feature fusion network, utilizing the attention weights between each pair of 3D channels, spatial widths, and spatial heights to improve efficiency by capturing the important features in three dimensions, thus enhancing the algorithm’s feature fusion capability. The WIoU loss function was employed to optimize the original bounding box loss function, enhancing the algorithm’s ability to locate the detected targets through a situation-compliant gradient gain allocation strategy. Additionally, a P2 small-target detection layer was added to strengthen the fusion of the deep and shallow feature information, thereby enhancing the algorithm’s ability to detect small-target defects. Experimental comparisons were conducted using the O-Rings dataset. The improved algorithm was compared with the YOLOv7-tiny algorithm, resulting in a 7.8% improvement in mAP and achieving a detection accuracy of 90.9%, meeting the needs of actual industrial production.

Key words: YOLOv7-tiny, rubber seal ring, defect detection, attention mechanism, small target detection layer

中图分类号:

TP391

张相胜, 杨骁. 基于改进YOLOv7-tiny的橡胶密封圈缺陷检测方法[J]. 图学学报, 2024, 45(3): 446-453.

ZHANG Xiangsheng, YANG Xiao. Defect detection method of rubber seal ring based on improved YOLOv7-tiny[J]. Journal of Graphics, 2024, 45(3): 446-453.

图/表 12

图1 YOLOv7-tiny算法网络结构图

Fig. 1 Network structure of the YOLOv7-tiny algorithm

图2 PConv与常规卷积结构图((a)常规卷积；(b)部分卷积)

Fig. 2 PConv and conventional convolutional structures ((a) Regular convolution; (b) Partial convolution (PConv))

图3 ELAN-P结构图

Fig. 3 ELAN-P structure

图4 GAM结构图

Fig. 4 GAM structure

图5 通道注意力子模块

Fig. 5 Channel attention submodule

图6 空间注意力子模块

Fig. 6 Space attention submodule

图7 改进的YOLOv7-tiny网络结构

Fig. 7 Improved YOLOv7-tiny network structure

图8 橡胶密封圈缺陷数据集样本图((a)表面粗糙；(b)划痕；(c)断裂；(d)缺料)

Fig. 8 Sample plot of rubber sealing ring defect data set ((a) Coarse; (b) Nick; (c) Crack; (d) Damage)

表1 橡胶密封圈缺陷数据集

Table 1 Rubber seal defect data set

缺陷类型	标签	数量/张
划痕	Nick	3 620
断裂	Crack	3 490
缺料	Damage	4 250
表面粗糙	Coarse	3 080

表2 消融实验结果对比

Table 2 Comparison of ablation experiment results

组别	Model	P/%	R/%	F1-score	mAP@0.5/%	Params/M	Time/ms
1组	YOLOv7-tiny	77.6	82.8	80.1	83.1	6.02	16.7
2组	YOLOv7-tiny+A	78.8	83.8	81.2	84.5	4.74	15.5
3组	YOLOv7-tiny+A+B	84.3	85.4	84.8	88.8	6.38	17.1
4组	YOLOv7-tiny+A+B+C	86.4	90.1	88.2	89.5	6.38	17.3
5组	YOLOv7-tiny+A+B+C+D	93.1	88.3	90.6	90.9	6.79	18.5

表3 对比实验结果对比

Table 3 Comparison of results of contrasting experiments

Model	mAP@0.5/%	Params/M	Time/ms
SSD	67.1	24.48	49.9
Faster RCNN	81.5	72.02	38.1
YOLOv5s	72.2	7.20	29.9
文献[11]	82.2	23.90	23.3
YOLOX-tiny	79.2	5.19	18.5
YOLOv6s	78.8	17.20	25.0
YOLOv7	88.9	37.20	40.4
文献[19]	89.1	42.80	40.5
YOLOv7-tiny	82.9	6.02	16.7
YOLOv8s	84.8	11.10	16.2
本文算法	90.9	6.79	18.5

图9 橡胶密封圈缺陷检测效果对比图((a)划伤；(b)缺料；(c)粗糙；(d)断裂)

Fig. 9 Comparison of the effect of detecting defects in rubber seals ((a) Nick; (b) Damage; (c) Coarse; (d) Crack)

参考文献 19

[1]	PENG G L, ZHANG Z J, LI W Q. Computer vision algorithm for measurement and inspection of O-rings[J]. Measurement, 2016, 94: 828-836.
[2]	DIESING G. How AI and machine vision intersect[J]. Quality, 2022, 61(2): 14.
[3]	YU Y W, YIN G F, DU L Q. Image classification for steel strip surface defects based on support vector machines[J]. Advanced Materials Research, 2011, 217-218: 336-340.
[4]	黄连, 刘晓军, 雷自力, 等. 基于奇异值分解的橡胶密封圈表面缺陷检测方法[J]. 润滑与密封, 2021, 46(11): 84-88. DOI
	HUANG L, LIU X J, LEI Z L, et al. Surface defect detection method of rubber sealing ring based on singular value decomposition[J]. Lubrication Engineering, 2021, 46(11): 84-88 (in Chinese). DOI
[5]	陶显, 侯伟, 徐德. 基于深度学习的表面缺陷检测方法综述[J]. 自动化学报, 2021, 47(5): 1017-1034.
	TAO X, HOU W, XU D. A survey of surface defect detection methods based on deep learning[J]. Acta Automatica Sinica, 2021, 47(5): 1017-1034 (in Chinese).
[6]	陶晓天, 何博侠, 张鹏辉, 等. 基于深度学习的航天密封圈表面缺陷检测[J]. 仪器仪表学报, 2021, 42(1): 199-206.
	TAO X T, HE B X, ZHANG P H, et al. Surface defect detection of aerospace sealing rings based on deep learning[J]. Chinese Journal of Scientific Instrument, 2021, 42(1): 199-206 (in Chinese).
[7]	李洪安, 郑峭雪, 陶若霖, 等. 基于深度学习的图像超分辨率研究综述[J]. 图学学报, 2023, 44(1): 1-15. DOI
	LI H A, ZHENG Q X, TAO R L, et al. Review of image super-resolutionbased on deep learning[J]. Journal of Graphics, 2023, 44(1): 1-15 (in Chinese).
[8]	REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2016: 779-788.
[9]	REDMON J, FARHADI A. Yolov3: an incremental improvement[EB/OL]. [2023-05-17]. http://arxivorg.abs.180402767.
[10]	KUZNETSOVA A, MALEVA T, SOLOVIEV V. Detecting apples in orchards using YOLOv3 and YOLOv5 in general and close-up images[C]// International Symposium on Neural Networks. Cham: Springer, 2020: 233-243.
[11]	朱文博, 夏林聪, 陈龙, 等. 基于改进YOLOv5的O型密封圈缺陷检测方法[J]. 上海理工大学学报, 2022, 44(5): 440-448.
	ZHU W B, XIA L C, CHEN L, et al. Defect detection method of O-ring based on improved YOLOv5[J]. Journal of University of Shanghai for Science and Technology, 2022, 44(5): 440-448 (in Chinese).
[12]	胡欣, 周运强, 肖剑, 等. 基于改进YOLOv5的螺纹钢表面缺陷检测[J]. 图学学报, 2023, 44(3): 427-437. DOI
	HU X, ZHOU Y Q, XIAO J, et al. Surface defect detection of threaded steel based on improved YOLOv5[J]. Journal of Graphics, 2023, 44(3): 427-437 (in Chinese). DOI
[13]	WANG C Y, BOCHKOVSKIY A, LIAO H Y M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2023: 7464-7475.
[14]	CHEN J R, KAO S H, HE H, et al. Run, don’t walk: chasing higher FLOPS for faster neural networks[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2023: 12021-12031.
[15]	LIU Y, SHAO Z, HOFFMANN N. Global attention mechanism: retain information to enhance channel-spatial interactions[EB/OL]. [2023-05-07]. http://arxivorg.abs.211205561.
[16]	TONG Z, CHEN Y U, XU Z. Wise-IoU: bounding box regression loss with dynamic focusing mechanism[EB/OL]. [2023-05-07]. http://arxivorg.abs.230110051.
[17]	LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot MultiBox detector[C]// European Conference on Computer Vision. Cham: Springer, 2016: 21-37.
[18]	REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[C]// IEEE Transactions on Pattern Analysis and Machine Intelligence. New York: IEEE Press, 2017: 1137-1149.
[19]	戚玲珑, 高建瓴. 基于改进YOLOv7的小目标检测[J]. 计算机工程, 2023, 49(1): 41-48. DOI
	QI L L, GAO J L. Small object detection based on improved YOLOv7[J]. Computer Engineering, 2023, 49(1): 41-48 (in Chinese). DOI