[1]刘玉洁,金 钧.基于改进的 YOLOv5s 绝缘子故障识别方法[J].机械与电子,2024,42(12):31-36.
　LIU Yujie,JIN Jun.Insulator Fault Identification Method of Overhead Contact System Based on Improved YOLOv5s[J].Machinery & Electronics,2024,42(12):31-36.
点击复制

基于改进的 YOLOv5s 绝缘子故障识别方法()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
42
期数:
2024年12期
页码:
31-36
栏目:
自动控制与检测
出版日期:
2024-12-24

文章信息/Info

Title:
Insulator Fault Identification Method of Overhead Contact System Based on Improved YOLOv5s
文章编号:
1001-2257 ( 2024 ) 12-0031-06
作者:
刘玉洁; 金 钧
大连交通大学自动化与电气工程学院,辽宁 大连 116028
Author(s):
LIU Yujie ;  JIN Jun
( School of Automation and Electrical Engineering , Dalian Jiaotong University , Dalian 116028 , China )
关键词:
绝缘子故障识别;  YOLOv5s 网络; AFPN ;  MPDIoU 损失函数
Keywords:
insulator fault recognition ;  YOLOv5s network ;  AFPN ;  MPDIoU loss function
分类号:
TP391.4
文献标志码:
A
摘要:
为解决高速铁路绝缘子故障检测中常见的错检、漏检等问题,以YOLOv5s 算法为基础进行优化提出 TASM-YOLOv5 算法。首先,增加 Triplet 注意力机制,以提升算法的特征提取能力;其次,引入 AFPN 渐进特征金字塔网络来提高特征融合利用能力,并且选用 SiLU 控制激活函数以提高稳定性;最后,更换损失函数为 MPDIoU 损失函数,可实现准确有效的边界框回归。实验结果表明, TASM-YOLOv5 算法的平均准确率较高,所得权重文件大小符合轻量化的要求,能有效提高绝缘子故障检测的精度。
Abstract:
In order to solve the? common problems such as wrong detection and missing detection in the fault detection of high speed railway insulators , the TASM-YOLOv5 algorithm is optimized based on the YOLOv5s algorithm.Firstly , the Triplet attention mechanism is added to improve the feature extraction capability of the algorithm.Secondly , the AFPN progressive feature pyramid network is introduced to improve the feature fusion utilization capability.SiLU control activation function is used to improve the stability.Finally , replacing the loss function with MPDIoU loss function can achieve accurate and efficient bounding box regression.The experimental results show that the average accuracy of the TASM-YOLOv5 algorithm is high , the weight file size meets the requirements of lightweight , and can effectively improve the accuracy of insulator fault detection.

参考文献/References:

[ 1 ] 党宏社,薛萌,郭琴 . 基于改进的 YOLOv4 绝缘子掉片故障检测方法[ J ] . 电瓷避雷器, 2022 ( 1 ): 211-218.

[ 2 ] 张长乐,金钧 . 基于 YOLOv5s 的绝缘子故障检测仿真研究[ J ] . 电子设计工程, 2024 , 32 ( 5 ): 178-182.
[ 3 ] 伍箴燎,吴正平,孙水发 . 基于改进 YOLOv5 算法的绝缘子多缺陷检测[ J / OL ] . 高压电器, 2024 : 1-11 [ 2024-07-03 ] .http : ∥kns.cnki.net / kcms / detail / 61.1127.TM.20230814.1815.002.html.
[ 4 ] 胡祥,李英娜 . 基于 YOLOv5 和 U Net++ 改进的绝缘子自爆故障检测方法[ J ] . 陕西理工大学学报(自然科学版),2022 , 38 ( 3 ): 22-29.
[ 5 ] 王海群,王康 . 基于 DMSANet-YOLOv7 的雾霾下绝缘子缺陷实时检测方法[ J ] . 无线电工程, 2024 , 54 ( 6 ):1431-1439.
[ 6 ] 邬开俊,徐泽浩,单宏全 . 基于 FasterNet 和 YOLOv5改进的玻璃绝缘子自爆缺陷快速检测方法[ J ] . 高电压技术,2024 , 50 ( 5 ): 1865-1876.
[ 7 ] MISRA D , NALAMADA T, ARASANIPALAI A U , et al.Rotate to attend : convolutional triplet attention module [ C ] ∥2021 IEEE Winter Conference on Applications of Computer Vision ( WACV ),2021 : 3138-3147.
[ 8 ] ZHAO Z B , PAN Y T , GUO G X , et al.YOLO-AFPN : Marrying YOLO and AFPN for external damage detection of transmission lines [ J ] .IET Generation , transmission and distribution , 2024 , 18 ( 9 ): 1935-1946.
[ 9 ] NISHIYAMA T , KUMAGAI A , KAMIYA K , et al. SILU : Strategy involving large-scale unlabeled logs for improving malware detector [ C ] ∥2020 IEEE Symposium on Computers and Communications ( ISCC ),2020 : 1-7.
[ 10 ] LI X , LI G , ZHANG Z.A study on the performance improvement of a conical bucket detection algorithm based on YOLOv8s [ J ] .World electric vehicle journal , 2024 , 15 ( 6 ): 238.
[ 11 ] 王爱丽,刘美红,薛冬,等 . 结合动态卷积和三重注意力机制的高光谱图像分类[ J ] . 激光与光电子学进展,2022 , 59 ( 10 ): 341-351.
[ 12 ] 吴慧海,沈文忠 . 基于 TA-YOLO 的电力设备红外图像检测方法[ J ] . 信息技术与信息化, 2022 ( 3 ): 17-20.
[ 13 ] 郭平秀,李启南,杨忠鹏 . 一种图像增强及改进海洋生物图像检测算法[ J ] . 计 算机工程与应用,2023 , 59( 8 ): 208-216.
[ 14 ] DAUBECHIES I , DEVORE R , FOUCART S , et al. Nonlinear approximation and ( deep ) ReLU networks [ J ] .Constructive approximation , 2022 , 55 ( 1 ): 127-172.
[ 15 ] ZHENG Z H , WANG P , LIU W , et al.Distance-IoU loss : faster and better learning for bounding box regression [ J ] .Proceedings of the AAAI conference on artificial intelligence , 2020 , 34 ( 7 ): 12993-13000.
[ 16 ] ZHENGZ H , WANG P , REN D W , et al.Enhancing geometric factors in model learning and inference for object detection and instance segmentation [ J ] .IEEE Transactions on cybernetics , 2022 , 52 ( 8 ): 8574-8586.
[ 17 ] REZATOFIGHI H, TSOI N , GWAK J Y , et al.Generalized intersection over union : a metric and a loss for bounding box regression [ C ] ∥2019 IEEE / CVF Conference on Computer Vision and Pattern Recognition ( CVPR ), 2019 : 658-666.
[ 18 ] 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.
[ 19 ] XU Y M , DONG D Z , WANG D S , et al.SSD-SGD : communication sparsification for distributed deep learning training [ J ] .ACM Transactions on architecture and code optimization , 2022 , 20 ( 1 ): 1-25.
[ 20 ] REDMON J , FARHADI A.YOLO9000 : better , faster , stronger [ C ] ∥2017 IEEE Conference on Computer Vision and Pattern Recognition ( CVPR ), 2017 : 6517-6525.
[ 21 ] 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 ( CVPR ), 2023 : 7464-7475.

备注/Memo

备注/Memo:
收稿日期: 2024-06-24
作者简介:刘玉洁 ( 1999- ),女,山东临沂人,硕士研究生,研究方向为电力系统故障识别;金 钧 ( 1970- ),男,辽宁大连人,博士,副教授,研究方向为轨道交通电气化与自动化。
更新日期/Last Update: 2025-01-07