基于自适应感受野的电力设备表面缺陷检测方法

doi:10.16182/j.issn1004731x.joss.22-0278

系统仿真学报 ›› 2023, Vol. 35 ›› Issue (7): 1572-1580.doi: 10.16182/j.issn1004731x.joss.22-0278

基于自适应感受野的电力设备表面缺陷检测方法

于豪¹(), 蒋锦霞², 赖晓翰², 梅峰², 王庆¹()

^1.西北工业大学计算机学院，陕西西安 710072
^2.国网浙江省电力有限公司信息通信分公司，浙江杭州 310020

收稿日期:2022-03-28 修回日期:2022-07-12 出版日期:2023-07-29 发布日期:2023-07-19
通讯作者: 王庆 E-mail:yh97@mail.nwpu.edu.cn;qwang@nwpu.edu.cn
作者简介:于豪(1998-)，男，硕士生，研究方向为计算机视觉和机器学习。E-mail：yh97@mail.nwpu.edu.cn
基金资助:
国家电网公司总部科技项目(5700-202019186A-0-0-00)

Surface Defect Detection of Power Equipment Using Adaptive Receptive Field Network

Hao Yu¹(), Jinxia Jiang², Xiaohan Lai², Feng Mei², Qing Wang¹()

^1.School of Computer Science, Northwestern Polytechnical University, Xi'an 710072, China
^2.Information & Telecommunication Branch, State Grid Zhejiang Electric Power Co. , Ltd. , Hangzhou 310020, China

Received:2022-03-28 Revised:2022-07-12 Online:2023-07-29 Published:2023-07-19
Contact: Qing Wang E-mail:yh97@mail.nwpu.edu.cn;qwang@nwpu.edu.cn

摘要/Abstract

摘要：

针对变电站电力设备覆冰、锈蚀、污秽等缺陷检测问题，提出了一种新的自适应感受野网络，其中结合注意力机制的自适应感受野模块可对多尺度特征进行有效融合。考虑到缺陷检测的小样本学习属性，还提出了一种基于真实纹理的电力设备表面缺陷仿真数据合成方法。在仿真数据集上的实验结果表明，该网络对跨设备表面缺陷的检测精度较高，同时具有体积小、运算速度快等优点。

关键词: 表面缺陷检测, 自适应感受野, 注意力机制, 多尺度特征, 仿真数据合成

Abstract:

For the detection of defects such as icing, rust, and contamination of power equipment in substations, a novel adaptive receptive field network (ARFN) is proposed,in which an adaptive receptive field module (ARFM) combined with the attention mechanism can effectively fuse multi-scale features. Considering the small sample learning attribute of defect detection, a power equipment surface defect simulation data synthesis method based on real texture is also proposed. The experimental results on the simulation dataset show that the network has high detection accuracy for surface defects across devices, while having advantages such as small size and fast operation speed.

Key words: surface defect detection, adaptive receptive field, attention mechanism, multi-scale feature, simulation data synthesis

中图分类号:

TP391.9

于豪, 蒋锦霞, 赖晓翰, 梅峰, 王庆. 基于自适应感受野的电力设备表面缺陷检测方法[J]. 系统仿真学报, 2023, 35(7): 1572-1580.

Hao Yu, Jinxia Jiang, Xiaohan Lai, Feng Mei, Qing Wang. Surface Defect Detection of Power Equipment Using Adaptive Receptive Field Network[J]. Journal of System Simulation, 2023, 35(7): 1572-1580.

图/表 6

图1

表1

网络参数

网络层次	ARFN-small	ARFN-medium	ARFN-large
FC, Softmax
Conv1	7×7, 16, s=2	7×7, 32, s=2	7×7, 64, s=2
Pooling	3×3, max pooling, s=2
Layer1	$1 × 1 161 3 × 3 46 1 × 1 161$	$1 × 1 321 3 × 3 86 1 × 1 321$	$1 × 1 641 3 × 3 166 1 × 1 641$
Layer2	$1 × 1 321 3 × 3 86 1 × 1 321$	$1 × 1 641 3 × 3 166 1 × 1 641$	$1 × 1 1281 3 × 3 326 1 × 1 1281$
Layer3	$1 × 1 641 3 × 3 166 1 × 1 641$	$1 × 1 1281 3 × 3 326 1 × 1 1281$	$1 × 1 2561 3 × 3 646 1 × 1 2561$
Layer4	$1 × 1 1281 3 × 3 326 1 × 1 1281$	$1 × 1 2561 3 × 3 646 1 × 1 2561$	$1 × 1 5121 3 × 3 1286 1 × 1 5121$

表1

表2

表3

图2

表4

参考文献 17

1	罗菁, 董婷婷, 宋丹, 等. 表面缺陷检测综述[J]. 计算机科学与探索, 2014, 8(9): 1041-1048.
	Luo Jing, Dong Tingting, Song Dan, et al. A Review on Surface Defect Detection[J]. Journal of Frontiers of Computer Science & Technology, 2014, 8(9): 1041-1048.
2	王伟旭, 刘志刚, 栗敏, 等. 基于曲线和点状奇异性的接触网用棒形瓷绝缘子外部缺陷的状态检测方法[J]. 电瓷避雷器, 2014(1): 16-22.
	Wang Weixu, Liu Zhigang, Li Min, et al. The State Detection Methods of External Defects for Catenary Rod Insulator Based on Curve and Punctiform Singularity[J]. Insulators and Surge Arresters, 2014(1): 16-22.
3	Yang Lin, Jiang Xiaolan, Hao Yanpeng, et al. Recognition of Natural Ice Types on In-service Glass Insulators Based on Texture Feature Descriptor[J]. IEEE Transactions on Dielectrics and Electrical Insulation, 2017, 24(1): 535-542.
4	Zuo Dan, Hu Hong, Qian Ronghui, et al. An Insulator Defect Detection Algorithm Based on Computer Vision[C]//2017 IEEE International Conference on Information and Automation (ICIA). Piscataway, NJ, USA:IEEE, 2017: 361-365.
5	Zhu Yongcan, Liu Cheng, Huang Xinbo, et al. Research on Image Recognition Method of Icicle Length and Bridging State on Power Insulators[J]. IEEE Access, 2019, 7: 183524-183531.
6	王红军, 黎邹邹, 邹湘军. 基于Adaboost与CNN的木材表面缺陷检测[J]. 系统仿真学报, 2019, 31(8):1636-1645.
	Wang Hongjun, Li Zouzou, Zou Xiangjun. Wood Surface Defect Detection Based on Adaboost and CNN[J]. Journal of System Simulation, 2019, 31(8): 1636-1645.
7	丘灵华, 朱铮涛. 基于深度学习的输电线路绝缘子缺陷检测研究[J]. 计算机应用研究, 2020, 37(增1): 358-360, 365.
	Qiu Linghua, Zhu Zhengtao. Research on Insulator Defect Detection of Transmission Line Based on Deep Learning[J]. Application Research of Computers, 2020, 37(S1): 358-360, 365.
8	Tao Xian, Zhang Dapeng, Wang Zihao, et al. Detection of Power Line Insulator Defects Using Aerial Images Analyzed With Convolutional Neural Networks[J]. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2020, 50(4): 1486-1498.
9	Kang Gaoqiang, Gao Shibin, Yu Long, et al. Deep Architecture for High-speed Railway Insulator Surface Defect Detection:Denoising Autoencoder With Multitask Learning[J]. IEEE Transactions on Instrumentation and Measurement, 2019, 68(8): 2679-2690.
10	Zhao Jian, You Zhenfei, Wang Xiaoyu, et al. A Hybrid CNN/Poisson Fusion Based Power Transformer External Defect Detecting Method[C]//2020 IEEE/IAS Industrial and Commercial Power System Asia (I&CPS Asia). Piscataway, NJ, USA:IEEE, 2020: 1453-1460.
11	Liu Wenqiang, Liu Zhigang, Wang Hui, et al. An Automated Defect Detection Approach for Catenary Rod-insulator Textured Surfaces Using Unsupervised Learning[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 69(10): 8411-8423.
12	Lin T Y, Dollár Piotr, Girshick R, et al. Feature Pyramid Networks for Object Detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ, USA: IEEE, 2017: 936-944.
13	Ren Shaoqing, He Kaiming, 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.
14	Girshick R. Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision (ICCV). Piscataway, NJ, USA: IEEE, 2015: 1440-1448.
15	Szegedy C, Liu Wei, Jia Yangqing, et al. Going Deeper With Convolutions[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ, USA: IEEE, 2015: 1-9.
16	Szegedy C, Vanhoucke V, Ioffe S, et al. Rethinking the Inception Architecture for Computer Vision[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ, USA: IEEE, 2016: 2818-2826.
17	Efros A A, Freeman W T. Image Quilting for Texture Synthesis and Transfer[C]//Proceedings of the 28th Annual Conference on Computer Graphics and Interactive Techniques. New York, NY, USA: Association for Computing Machinery, 2001: 341-346.

网络	覆冰	锈蚀	污秽	mAP
VGG16	0.171	0.133	0.177	0.160
ResNet50	0.909	0.908	0.906	0.907
ResNet101	0.909	0.908	0.908	0.908
ResNet50-fpn	0.914	0.913	0.913	0.913
ResNet50-afpn	0.914	0.914	0.913	0.914
ARFN-small	0.891	0.837	0.817	0.849
ARFN-medium	0.907	0.895	0.898	0.900
ARFN-large	0.909	0.908	0.907	0.908

网络	模型大小/MB	计算复杂度/Gflops	参数数量(M)
VGG16	163.63	192.79	24.22
ResNet50	182.23	264.49	28.30
ResNet101	327.35	300.62	47.29
ResNet50-fpn	217.32	320.29	31.53
ResNet50-afpn	200.78	323.67	32.44
ARFN-small	4.28	5.89	0.60
ARFN-medium	10.26	22.12	1.44
ARFN-large	28.56	85.65	3.97

缺陷类型	ARFN_s/w	ARFN_s/o
mAP	0.849	0.810
覆冰	0.891	0.873
锈蚀	0.837	0.796
污秽	0.817	0.760

[1]	魏赟, 罗琦, 赵迎志. 基于自适应融合和注意力细化的语义分割模型[J]. 系统仿真学报, 2023, 35(6): 1226-1234.
[2]	付玉, 张垚, 赵萌, 王绵沼, 郑江鹏, 贾晨, 陈胜勇. 基于仿真数据迁移学习的固定翼无人机检测[J]. 系统仿真学报, 2023, 35(5): 998-1007.
[3]	石鼎, 燕雪峰, 宫丽娜, 张静宣, 关东海, 魏明强. 强化学习驱动的海战场多智能体协同作战仿真算法[J]. 系统仿真学报, 2023, 35(4): 786-796.
[4]	向南, 王璐, 贾崇柳, 蹇越谋, 马小霞. 改进YOLO的遮挡行人检测仿真[J]. 系统仿真学报, 2023, 35(2): 286-299.
[5]	孙红, 张玉香, 凌岳览. 基于损失提取反馈注意网络的图像超分辨率重建研究[J]. 系统仿真学报, 2023, 35(2): 308-317.
[6]	金炜东, 张述礼, 唐鹏, 张曼. 基于稠密残差块与通道像素注意力的图像去雾网络[J]. 系统仿真学报, 2022, 34(8): 1663-1673.
[7]	邱俊杰, 郑红, 程云辉. 基于多尺度LSTM预测模型研究[J]. 系统仿真学报, 2022, 34(7): 1593-1604.
[8]	王银, 王飞翔, 孙前来. 多尺度特征融合车辆检测方法[J]. 系统仿真学报, 2022, 34(6): 1219-1229.
[9]	尹诗, 侯国莲, 迟岩, 弓林娟, 胡晓东. 风电机组发电机前轴承健康度预测方法及实现[J]. 系统仿真学报, 2021, 33(6): 1323-1333.
[10]	徐胜, 冯文宇, 刘志诚, 涂鑫涛, 费敏锐, 张堃. 基于机器视觉的复杂环境下精确手势识别算法研究[J]. 系统仿真学报, 2021, 33(10): 2460-2469.
[11]	孔亦思, 胡晓峰, 朱丰, 陶九阳. 战场态势感知中的注意力机制探析[J]. 系统仿真学报, 2017, 29(10): 2233-2241.

基于自适应感受野的电力设备表面缺陷检测方法

Surface Defect Detection of Power Equipment Using Adaptive Receptive Field Network

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 17

相关文章 11

编辑推荐

Metrics

本文评价