基于YOLOv5和DeepSORT的金刚石锯丝磨损监测

袁俊涛; 赵礼刚; 秦齐; 王天旭; 刘志强

doi:10.13394/j.cnki.jgszz.2022.0065

基于YOLOv5和DeepSORT的金刚石锯丝磨损监测

doi: 10.13394/j.cnki.jgszz.2022.0065

袁俊涛¹,
赵礼刚^{1, 2,},
秦齐¹,
王天旭¹,
刘志强^{1, 2}

1.
江苏科技大学机械工程学院, 江苏镇江 212003
2.
江苏科技大学, 江苏省船海机械装备先进制造重点实验室, 江苏镇江 212003

详细信息

作者简介:
赵礼刚，男，1978年生，博士，讲师。主要研究方向：精密加工与特种加工、机器视觉等。 E-mail：zhaoligang@just.edu.cn

中图分类号: TQ164；TG74；TP391
计量
- 文章访问数: 513
- HTML全文浏览量: 264
- PDF下载量: 48
- 被引次数: 0
出版历程
- 收稿日期: 2022-05-05
- 修回日期: 2022-08-05
- 录用日期: 2022-08-12
- 网络出版日期: 2023-01-04
- 刊出日期: 2023-02-20

Wear monitoring of diamond saw wire based on YOLOv5 and DeepSORT

YUAN Juntao¹,
ZHAO Ligang^{1, 2
,},
QIN Qi¹,
WANG Tianxu¹,
LIU Zhiqiang^{1, 2}

1.
School of Mechanical Engineering, Jiangsu University of Science and Technology, Zhenjiang 212003, Jiangsu, China
2.
Jiangsu Provincial Key Laboratory of Advanced Manufacturing for Marine Mechanical Equipment, Jiangsu University of Science and Technology, Zhenjiang 212003, Jiangsu, China

摘要

摘要:
为提高金刚石线锯切割的效率和质量，满足实时监测锯丝磨损的需求，提出一种基于改进的YOLOv5检测算法，在YOLOv5的基础上融合坐标注意力机制和BiFPN模块，使检测精确度、召回率、平均精度均值分别提高1.7%、3.7%、3.2%，能够有效检测不同磨损程度的磨粒；再连接DeepSORT多目标跟踪算法，设置虚拟检测线，统计不同磨损程度的磨粒数量，进而监测金刚石锯丝的磨损情况。
- 金刚石锯丝 /
- 目标检测 /
- YOLOv5 /
- DeepSORT
Abstract:
In order to improve the efficiency and quality of diamond wire saw cutting and meet the demand of real-time monitoring of saw wire wear, a detection algorithm based on improved YOLOv5 was proposed. The algorithm combined coordinate attention mechanism and BiFPN module on the basis of YOLOv5. The detection accuracy, recall rate and average accuracy were increased by 1.7%, 3.7% and 3.2% respectively. Abrasive particles with different wear degrees can be effectively detected. Besides, the DeepSORT multi-target tracking algorithm was connected to set up a virtual detection line, count the number of abrasive particles with different wear degrees, and monitor the wear of diamond saw wire.
- diamond saw wire /
- target detection /
- YOLOv5 /
- DeepSORT

HTML全文

图 1 拍摄设备布局

Figure 1. Layout of shooting equipment

下载: 全尺寸图片幻灯片

图 2 锯丝拍摄效果

Figure 2. Photo of the diamond wire saw

下载: 全尺寸图片幻灯片

图 3 磨粒分类

Figure 3. Classification of diamond abrasive

下载: 全尺寸图片幻灯片

图 4 检测效果对比

Figure 4. Comparison of detection effects

下载: 全尺寸图片幻灯片

图 5 磨粒数量统计界面

Figure 5. Diamond abrasive quantity statistics interface

下载: 全尺寸图片幻灯片

图 6 失效磨粒数

Figure 6. Number of useless abrasive

下载: 全尺寸图片幻灯片

图 7 不具备磨粒特征的锯丝

Figure 7. Saw wire without abrasive characteristics

下载: 全尺寸图片幻灯片

表 1 消融对比试验

Table 1. Ablation experiments

模型	CA	BiFPN	精确度 P_re / %	召回率 R_re / %	平均精度值 M_mAP / %	推理时间 t / (ms·帧⁻¹)
YOLOv5	×	×	79.8	81.1	84.8	8
模型1	√	×	81.2	83.2	87.5	8
模型2	×	√	80.4	83.6	87.0	9
最终模型	√	√	81.5	84.8	88.0	10

下载: 导出CSV

参考文献(15)

[1]	倪永明, 黎振, 李国和. 蓝宝石加工过程中金刚石线锯生命周期表面形貌及抗拉极限载荷实验研究 [J]. 金刚石与磨料磨具工程,2017,37(4):62-66. doi: 10.13394/j.cnki.jgszz.2017.4.0013 NI Yongming, LI Zhen, LI Guohe. Experimental study on life-cycle surface morphology and tensile limit load of diamond wire saw in sapphire machining [J]. Diamond & Abrasives Engineering,2017,37(4):62-66. doi: 10.13394/j.cnki.jgszz.2017.4.0013
[2]	黄波, 高玉飞, 葛培琪. 金刚石线锯切割单晶硅表面缺陷与锯丝磨损分析 [J]. 金刚石与磨料磨具工程,2011,31(1):53-57. doi: 10.3969/j.issn.1006-852X.2011.01.013 HUANG Bo, GAO Yufei, GE Peiqi. Surface sefects and wire wear analysis of diamond wire saw cutting monocrystalline silicon [J]. Diamond & Abrasives Engineering,2011,31(1):53-57. doi: 10.3969/j.issn.1006-852X.2011.01.013
[3]	GIRSHICK R, DONAHUE J, DARRELL T, et al. Proceedings of the IEEE conference on computer vision and pattern recognition, Columbus, June 23-28, 2014 [C]. Washington: IEEE Computer Society, 2014: 580-587.
[4]	HE K M, ZHANG X, REN S, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(9):1904-1916. doi: 10.1109/TPAMI.2015.2389824
[5]	REN S, HE K, GIRSHICK R, et al. Faster R-CNN: Towards real-time object detection with region proposal networks [J]. IEEE Transactions on Pattern Analysis & Machine Intelligence,2017,39(6):1137-1149.
[6]	REDMON J, DIVVALA S, GIRSHICK R, et al. Proceedings of the IEEE conference on computer vision and pattern recognition, Las Vegas, June 27-30, 2016 [C]. Washington: IEEE Computer Society, 2016: 779-788.
[7]	REDMON J, FARHADI A. Proceedings of the IEEE conference on computer vision and pattern recognition, Honolulu, July 21-26, 2017 [C]. Washington: IEEE Computer Society, 2017: 6517-6525.
[8]	REDMON J, FARHADI A. YOLOv3: An incremental improvement [J]. Computer Vision and Pattern Recognition, 2018, 276(7): 126-134.
[9]	LIU W, ANGUELOV D, ERHAN D, et al. European conference on computer vision, Amsterdam, October 11-14, 2016 [C]. Berlin, Heidelberg: Springer, 2016: 21-37.
[10]	张宸嘉, 朱磊, 俞璐. 卷积神经网络中的注意力机制综述 [J]. 计算机工程与应用,2021,57(20):64-72. ZHANG Chengjia, ZHU Lei, YU Lu. A review of attention mechanisms in convolutional neural networks [J]. Computer Engineering and Applications,2021,57(20):64-72.
[11]	HU J, SHEN L, SUN G. Proceedings of the IEEE conference on computer vision and pattern recognition, Salt Lake City, June 18-22, 2018 [C]. Washington: IEEE Computer Society, 2018: 7132-7141.
[12]	WANG Q, WU B, ZHU P, et al. IEEE conference on computer vision and pattern recognition, Seattle, June 13-19 2020 [C]. Washington: IEEE Computer Society, 2020: 11531-11539.
[13]	HOU Q B, ZHOU D Q, FENG J S. IEEE conference on computer vision and pattern recognition, Nashville, June 20-25, 2021 [C]. Washington: IEEE Computer Society, 2021: 13713-13722.
[14]	TAN M, PANG R, LE Q V. IEEE conference on computer vision and pattern recognition, Seattle, June 13-19 2020 [C]. Washington: IEEE Computer Society, 2020:10778-10787.
[15]	WOJKE N, BEWLEY A, PAULUS D. IEEE international conference on image processing, Beijing, September 17-20, 2017 [C]. Washington: IEEE Computer Society, 2018: 3645-3649.