基于自适应关键点的破损旋转绝缘子检测方法

doi:10.19678/j.issn.1000-3428.0065725

计算机工程 ›› 2023, Vol. 49 ›› Issue (9): 272-278. doi: 10.19678/j.issn.1000-3428.0065725

基于自适应关键点的破损旋转绝缘子检测方法

龙玉江¹, 卫薇¹, 舒彧¹, 张正刚², 王道累², 李峰²^,*

1. 贵州电网有限责任公司信息中心, 贵阳 550000
2. 上海电力大学电气工程学院, 上海 201306

收稿日期:2022-09-13 出版日期:2023-09-15 发布日期:2023-09-14
通讯作者: 李峰
作者简介:
龙玉江(1976—)，男，高级工程师，主研方向为信息技术
卫薇，高级工程师、硕士
舒彧，高级工程师、硕士
张正刚，硕士研究生
王道累，教授、博士
基金资助:
国家自然科学基金(61502297)

Detection Method for Damaged Rotating Insulator Based on Adaptive Key Points

Yujiang LONG¹, Wei WEI¹, Yu SHU¹, Zhenggang ZHANG², Daolei WANG², Feng LI²^,*

1. Information Center, Guizhou Power Grid Co., Ltd., Guiyang 550000, China
2. College of Electric Power Engineering, Shanghai University of Electric Power, Shanghai 201306, China

Received:2022-09-13 Online:2023-09-15 Published:2023-09-14
Contact: Feng LI

摘要/Abstract

摘要：

在输电线路检测图像中受绝缘子方向不确定、背景复杂、遮挡物等因素影响，导致漏检和定位准确率降低。提出一种基于自适应关键点的破损旋转绝缘子检测方法。将Oriented RepPoints目标检测算法作为Baseline并加以改进。针对Oriented RepPoints算法的方向信息感知较弱以及特征表示能力差的问题，结合递归特征金字塔和坐标注意力机制，构建递归强化特征金字塔网络(Re-FPN)。其中，CA模块从2个方向聚合特征，捕获方向感知和位置敏感信息，提升模型的定位和识别能力，Re-FPN在经典特征金字塔网络基础上进行递归式特征提取，将输入图像特征进行反复提取，使得目标检测的错误回传信息能够更直接反馈以调整主干网络参数，从而提升提取特征对检测的适用性。在自建绝缘子图像数据集上进行训练和测试，实验结果表明，该方法的绝缘子检测准确率达到96.1%，相比Oriented RepPoints算法提升2.1个百分点，同时该算法性能表现优于现有主流目标检测算法，具有较高的工程应用价值。

关键词: 绝缘子, 旋转框, 目标检测, 特征金字塔网络, 注意力机制

Abstract:

Aiming at the problems of missing detection and inaccurate positioning caused by uncertain insulator direction, complex background, and occlusion in transmission line detection images, this study proposes a detection method for damaged rotating insulator based on adaptive key points. The Oriented RepPoints target detection algorithm is selected as the baseline and improved. To solve the problems of weak directional information perception and poor feature representation in the Oriented RepPoints algorithm, this study proposes a Recursive-enhanced Feature Pyramid Network (Re-FPN) based on the Recursive Feature Pyramid (RFP) and Coordinate Attention (CA) mechanism. The CA module improves the positioning and recognition capability of the model by aggregating features from two directions to capture directional perception and location-sensitive information. The Re-FPN performs recursive feature extraction based on the classic FPN, repeatedly purifying the features of the input image so that the error feedback information of target detection can more directly feedback and adjust the parameters of the backbone, greatly improving the applicability of extracted features for detection. Training and testing are conducted on a self-built insulator image dataset. The experimental results showed that the accuracy of insulator detection by the proposed method reached 96.1%, which is 2.1 percentage points higher than that of the baseline Oriented RepPoints algorithm. At the same time, the improved algorithm performs better than the existing mainstream target detection and calculation methods and has a high engineering application value.

Key words: insulator, rotating frame, object detection, Feature Pyramid Network (FPN), attention mechanism

龙玉江, 卫薇, 舒彧, 张正刚, 王道累, 李峰. 基于自适应关键点的破损旋转绝缘子检测方法[J]. 计算机工程, 2023, 49(9): 272-278.

Yujiang LONG, Wei WEI, Yu SHU, Zhenggang ZHANG, Daolei WANG, Feng LI. Detection Method for Damaged Rotating Insulator Based on Adaptive Key Points[J]. Computer Engineering, 2023, 49(9): 272-278.

http://www.ecice06.com/CN/Y2023/V49/I9/272

图/表 9

图1 不同旋转框生成方法的提取效果对比

Fig.1 Extraction effect comparison among different rotation box generation methods

图2 Oriented RepPoints结构

Fig.2 Structure of Oriented RepPoints

图3 递归强化特征金字塔网络结构

Fig.3 Structure of recursive-enhanced feature pyramid network

图4 空洞空间金字塔池化网络结构

Fig.4 Structure of atrous spatial pyramid pooling network

图5 坐标注意力机制结构

Fig.5 Structure of coordinate attention mechanism

图6 YOLOv4算法与本文算法的可视化结果对比

Fig.6 Visual results comparison between YOLOv4 algorithm and the proposed algorithm

参考文献 25

1	MIAO X, LIU X, CHEN J, et al. Insulator detection in aerial images for transmission line inspection using single shot multibox detector. IEEE Access, 2019, 7, 9945- 9956. doi: 10.1109/ACCESS.2019.2891123
2	唐靓, 余明慧, 武明虎, 等. 基于改进YOLOv5的绝缘子缺陷检测算法. 华中师范大学学报(自然科学版), 2022, 56(5): 771- 780. URL
	TANG J, YU M H, WU M H, et al. Insulator defect detection algorithm based on improved YOLOv5. Journal of Central China Normal University (Natural Sciences), 2022, 56(5): 771- 780. URL
3	MENENDEZ O, AUAT CHEEIN F A, PEREZ M, et al. Robotics in power systems: enabling a more reliable and safe grid. IEEE Industrial Electronics Magazine, 2017, 11(2): 22- 34. doi: 10.1109/MIE.2017.2686458
4	王道累, 孙嘉珺, 张天宇, 等. 基于改进生成对抗网络的玻璃绝缘子自爆缺陷检测方法. 高电压技术, 2022, 48(3): 1096- 1103. URL
	WANG D L, SUN J J, ZHANG T Y, et al. Self-explosion defect detection method of glass insulator based on improved generative adversarial network. High Voltage Engineering, 2022, 48(3): 1096- 1103. URL
5	NEUBECK A, VAN GOOL L. Efficient non-maximum suppression[C]//Proceedings of the 18th International Conference on Pattern Recognition. Washington D. C., USA: IEEE Press, 2006: 850-855.
6	CHEN J W, XU X, DANG H S. Fault detection of insulators using second-order fully convolutional network model. Mathematical Problems in Engineering, 2019, 2019, 1- 10.
7	MA J Q, SHAO W Y, YE H, et al. Arbitrary-oriented scene text detection via rotation proposals. IEEE Transactions on Multimedia, 2018, 20(11): 3111- 3122. doi: 10.1109/TMM.2018.2818020
8	JIANG Y Y, ZHU X Y, WANG X B, et al. R²CNN: rotational region CNN for orientation robust scene text detection[EB/OL]. [2022-08-08]. https://arxiv.org/ftp/arxiv/papers/1706/1706.09579.pdf.
9	LIAO M H, SHI B G, BAI X. TextBoxes: a single-shot oriented scene text detector. IEEE Transactions on Image Processing, 2018, 27(8): 3676- 3690. doi: 10.1109/TIP.2018.2825107
10	ZHOU X Y, YAO C, WEN H, et al. EAST: an efficient and accurate scene text detector[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2017: 2642-2651.
11	ZHANG G J, LU S J, ZHANG W. CAD-net: a context-aware detection network for objects in remote sensing imagery. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(12): 10015- 10024. doi: 10.1109/TGRS.2019.2930982
12	AZIMI S M, VIG E, BAHMANYAR R, et al. Towards multi-class object detection in unconstrained remote sensing imagery[C]//Proceedings of the European Conference on Computer Vision. Berlin, Germany: Springer, 2019: 150-165.
13	YANG X, YANG J R, YAN J C, et al. SCRDet: towards more robust detection for small, cluttered and rotated objects[C]//Proceedings of International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2020: 8231-8240.
14	XU Y C, FU M T, WANG Q M, et al. Gliding vertex on the horizontal bounding box for multi-oriented object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43(4): 1452- 1459. doi: 10.1109/TPAMI.2020.2974745
15	WEI H R, ZHANG Y, CHANG Z H, et al. Oriented objects as pairs of middle lines[EB/OL]. [2022-08-08]. https://arxiv.org/abs/1912.10694v3.
16	LI Y Y, HUANG Q, PEI X A, et al. RADet: refine feature pyramid network and multi-layer attention network for arbitrary-oriented object detection of remote sensing images. Remote Sensing, 2020, 12(3): 389. doi: 10.3390/rs12030389
17	WANG J W, DING J A, GUO H W, et al. Mask OBB: a semantic attention-based mask oriented bounding box representation for multi-category object detection in aerial images. Remote Sensing, 2019, 11(24): 2930. doi: 10.3390/rs11242930
18	SHAO Z, ZHANG X, ZHANG T, et al. RBFA-Net: a rotated balanced feature-aligned network for rotated SAR ship detection and classification. Remote Sensing, 2022, 14(14): 3345. doi: 10.3390/rs14143345
19	CHAN S X, ZHENG J C, WANG L. et al. Rotating object detection in remote-sensing environment. Soft Computing, 2022, 26, 8037- 8045. doi: 10.1007/s00500-022-07058-z
20	LI W T, CHEN Y J, HU K X, et al. Oriented RepPoints for aerial object detection. [EB/OL]. [2022-08-08]. https://arxiv.org/abs/2105.11111, 25 Nov 2021.
21	HOU Q B, ZHOU D Q, FENG J S. Coordinate attention for efficient mobile network design[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2021: 13708-13717.
22	TAO X, ZHANG D P, WANG Z H, et al. Detection of power line insulator defects using aerial images analyzed with convolutional neural networks. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2020, 50(4): 1486- 1498. doi: 10.1109/TSMC.2018.2871750
23	TAN M X, PANG R M, LE Q V. EfficientDet: scalable and efficient object detection[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2020: 10778-10787.
24	DUAN K W, BAI S, XIE L X, et al. CenterNet: keypoint triplets for object detection[EB/OL]. [2022-08-08]. https://arxiv.org/abs/1904.08189.
25	BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: optimal speed and accuracy of object detection [EB/OL]. [2022-08-08]. https://arxiv.org/abs/2004.10934.

[1]	杨静, 陆铭华, 马洁琼, 吴金平, 刘星璇. 基于交替循环神经网络的水下防御态势预测方法[J]. 计算机工程, 2023, 49(9): 69-78.
[2]	孙龙, 张荣芬, 刘宇红, 饶庭漓. 监控视角下密集人群口罩佩戴检测算法[J]. 计算机工程, 2023, 49(9): 313-320.
[3]	李嘉新, 侯进, 盛博莹, 周宇航. 基于改进YOLOv5的遥感小目标检测网络[J]. 计算机工程, 2023, 49(9): 256-264.
[4]	苏晓东, 李世洲, 赵佳圆, 亮洪宇, 张玉荣, 徐红岩. 基于多级叠加和注意力机制的图像语义分割[J]. 计算机工程, 2023, 49(9): 265-271, 278.
[5]	韩璐, 霍纬纲, 张永会, 刘涛. 基于多尺度特征融合与双注意力机制的多元时间序列预测[J]. 计算机工程, 2023, 49(9): 99-108.
[6]	丰芳宇, 罗晓曙, 蒙志明, 王广宇. 基于抗混叠残差注意力网络的人脸表情识别[J]. 计算机工程, 2023, 49(8): 190-198.
[7]	徐春波, 闫娟, 杨慧斌, 王博, 吴晗. 基于目标检测和语义分割的视觉SLAM算法[J]. 计算机工程, 2023, 49(8): 199-206, 214.
[8]	宋志娜, 李莎, 杨建明, 徐川. 基于特征与区域定位增强的遥感舰船目标检测[J]. 计算机工程, 2023, 49(8): 257-264.
[9]	王书朋, 何引弟. 融合特征注意力机制的非均匀光照图像增强算法[J]. 计算机工程, 2023, 49(8): 232-239.
[10]	刘昊鑫, 董超, 勾智楠, 高凯. 融合混合表征的小样本关系抽取方法[J]. 计算机工程, 2023, 49(8): 63-68.
[11]	杨长沛, 廖列法. 基于门控空洞卷积特征融合的中文命名实体识别[J]. 计算机工程, 2023, 49(8): 85-95.
[12]	刘俊豪, 王美林, 谢兴, 宋烨兴, 许莉花. 基于改进YOLOv5的皮革瑕疵检测算法[J]. 计算机工程, 2023, 49(8): 240-249.
[13]	马娜, 温廷新, 贾旭, 李晓会. 复杂光照条件下自适应的车脸重识别模型[J]. 计算机工程, 2023, 49(8): 275-282, 290.
[14]	陈露萌, 曹彦彦, 黄民, 谢鑫钢. 基于改进YOLOv5的火焰检测方法[J]. 计算机工程, 2023, 49(8): 291-301, 309.
[15]	李强龙, 周新文, 位梦恩, 甘阳洲. 基于条形池化和注意力机制的街道场景红外目标检测算法[J]. 计算机工程, 2023, 49(8): 310-320.

选择文件类型/文献管理软件名称

选择包含的内容

基于自适应关键点的破损旋转绝缘子检测方法

Detection Method for Damaged Rotating Insulator Based on Adaptive Key Points

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 25

相关文章 15

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于自适应关键点的破损旋转绝缘子检测方法

Detection Method for Damaged Rotating Insulator Based on Adaptive Key Points

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 25

相关文章 15

编辑推荐

Metrics

本文评价