改进YOLO v7的绝缘子检测与定位

doi:10.3778/j.issn.1002-8331.2306-0094

摘要/Abstract

摘要： 针对电力系统目标检测任务中绝缘子大小不一、背景干扰等影响而导致精度低、漏检率高的问题，提出了基于改进YOLO v7绝缘子检测与定位方法。在YOLO v7骨干网络中加入轻量级注意力机制（convolutional block attention module, CBAM），使网络模型从通道、空间两个方面更加关注绝缘子特征，降低绝缘子检测中的漏检率。在网络模型深层添加集中特征金字塔（concentrated feature pyramid, CFP），使不同尺度的特征图进行信息交换和聚合，进而获得更加全面的绝缘子特征，提高绝缘子检测精度。通过k-means算法对预选框聚类，得到最适合绝缘子预选框大小。实验结果表明，改进以后的YOLO v7网络模型平均检测精度（mean average precision, mAP）达到96.2%，精准率为90.8%，召回率为93.8%。改进的方法在电力系统绝缘子检测中具有较广泛的应用前景。

关键词: 目标检测, 深度学习, 无人机巡检图像, 绝缘子识别

Abstract: This paper aims to address the problems of low accuracy and high leakage rate due to the influence of different insulator sizes and background interference in the target detection task of power systems. Firstly, a convolutional block attention module (CBAM) is added to the YOLO v7 backbone network to make the network model pay more attention to the insulator features from both channel and space aspects and reduce the leakage rate in insulator detection. Secondly, a concentrated feature pyramid (CFP) is added to the deeper layer of the network model to allow the information exchange and aggregation of feature maps at different scales, thus obtaining more comprehensive insulator features and improving insulator detection accuracy. Finally, the k-means algorithm is used to cluster the preselected frames to obtain the most suitable insulator preselected frame size. The experimental results show that the improved YOLO v7 network model has a detection mAP (mean average precision) of 96.2%, a precision of 90.8%, and a recall of 93.8%. The improved method in this paper has a wide application prospect in the insulator detection of power systems.

Key words: object detection, deep learning, UAV patrol image, insulator identification

张剑锐, 魏霞, 张林鍹, 陈燕楠, 卢杰. 改进YOLO v7的绝缘子检测与定位[J]. 计算机工程与应用, 2024, 60(4): 183-191.

ZHANG Jianrui, WEI Xia, ZHANG Linxuan, CHEN Yannan, LU Jie. Improving Detection and Positioning of Insulators in YOLO v7[J]. Computer Engineering and Applications, 2024, 60(4): 183-191.

参考文献

[1] 王德文, 李业东. 基于WGAN图片去模糊的绝缘子目标检测[J]. 电力系统保护与控制, 2020, 40(5): 188-198.
WANG D W, LI Y D. Insulator target detection based on WGAN image deblurring[J]. Electric Power Automation Equipment, 2020, 40(5): 188-198.
[2] 汤浩征. 基于绝缘子污闪泄露电流的预警系统设计与实现[D]. 北京: 华北电力大学(北京), 2017.
TANG H Z. Design and implementation of early warning system based on insulator pollution flash over leakage current[D]. Beijing: North China Electric Power University (Beijing), 2017.
[3] CHEN L, LIN F, CHEN M J, et al. Influence for ambient relative humidity and pollution on infrared detection of zero resistance insulators[J]. Frontiers in Energy Research, 2022. DOI: 10.3389/fenrg.2022.942408.
[4] 王维, 鲁海亮, 李纯, 等. 基于不同倾角下水珠形态量化特征的V串复合绝缘子憎水性判定方法[J]. 高电压技术, 2023, 49(4): 1665-1674.
WANG W, LU H L, LI C, et al. Determination method of hydrophobicity of V-string composite insulators based on quantitative characteristics of water droplet morphology under different inclination angles[J]. High Voltage Engineering, 2023, 49(4): 1665-1674.
[5] 骆国防, 陆克昀, 王月强. 特高压变电站绝缘子超声波检测[J]. 华东电力, 2014, 42(12): 2818-2820.
LUO G F, LU K Y, WANG Y Q. Ultrasonic testing of UHV substation insulator[J]. East China Electric Power, 2014, 42(12): 2818-2820.
[6] 梁捷生. 复合绝缘子的表面电场仿真与泄露电流实验研究[J]. 能源与环境, 2022(1): 24-27.
LIANG J S. Simulation of surface electric field and experimental study of leakage current in composite insulators[J]. Energy and Environment, 2022(1): 24-27.
[7] 蔡正梓, 李卫国, 曹文彬, 等. 基于WiFi通信的绝缘子电阻检测机器人的研制[J]. 中国电力, 2015, 48(12): 23-26.
CAI Z X, LI W G, CAO W B, et al. Development of an insulator resistance testing robot based on WiFi communication[J]. Electric Power, 2015, 48(12): 23-26.
[8] 董军, 李博, 翟洪达, 等. 基于多旋翼无人机的复合绝缘子憎水性带电检测装置研究[J]. 电力科学与技术学报, 2020, 35(6): 171-179.
DONG J, LI B, ZHAI H D, et al. Research on the hydrophobic testing device of composite insulator based on the multi-rotor UAV[J]. Journal of Electric Power Science and Technology, 2020, 35(6): 171-179.
[9] 师涛. 悬式绝缘子内部人工缺陷的相控阵超声波检测策略研究[J]. 电子制作, 2019(16): 90-92.
SHI T. Phased-array ultrasound of artificial defects inside suspension insulators detection strategy study[J]. Practical Electronics, 2019(16): 90-92.
[10] 刘子英, 肖建华, 邓芳明, 等. 基于可见光图像识别的绝缘子污秽等级判别[J]. 传感器与微系统, 2019, 38(12): 136-139.
LIU Z Y, XIAO J H, DENG F M, et al. Identification of insulator contamination grade based on visible light image recognition[J]. Transducer and Microsystem Technologies, 2019, 38(12): 136-139.
[11] 王晶晶, 杨俊杰. 改进Canny算法在绝缘子串红外图像边缘检测中的应用[J]. 电瓷避雷器, 2022(3): 179-185.
WANG J J, YANG J J. Application of improved Canny algorithm in edge detection of high voltage insulator string[J]. Insulator and Surge Arresters, 2022(3): 179-185.
[12] 刘开培, 李博强, 秦亮, 等. 深度学习目标检测算法在架空输电线路绝缘子缺陷检测中的应用研究综述[J]. 高电压技术, 2023, 49(9): 3584-3595.
LIU T P, LI B Q, QIN L, et al. Review of application research of deep learning object detection algorithms in insulator defect detection of overhead transmission lines[J]. High Voltage Engineering, 2023, 49(9): 3584-3595.
[13] 孙磊, 吴文海, 柯坚, 等. 基于可见光图像的铁路接触网绝缘子污秽状况检测[J]. 电瓷避雷器, 2019(5): 240-244.
SUN L, WU W H, KE J, et al. Detection of contamination of railway catenary insulator based on visible light image[J]. Insulators and Surge Arresters, 2019(5): 240-244.
[14] 高嵩, 陆倚鹏, 王笑倩, 等. 基于深度学习的悬式瓷绝缘子红外图像识别方法[J]. 电力科学与技术学报, 2020, 35(5): 119-125.
GAO S, LU Y P, WANG X Q, et al. Infrared image recognition method of porcelain disc-suspended insulators based on deep learning technology[J]. Journal of Electric Power Science and Technology, 2020, 35(5): 119-125.
[15] 刘传洋, 吴一全. 基于深度学习的输电线路视觉检测方法研究进展[J]. 中国电机工程学报, 2023, 43(19): 7423-7446.
LIU C Y, WU Y Q. Research progress of vision detection methods based on deep learning for transmission line[J]. Proceedings of the CSEE, 2023, 43(19): 7423-7446.
[16] 林刚, 王波, 彭辉, 等. 基于改进Faster-RCNN的输电线巡检图像多目标检测及定位[J]. 电力自动化设备, 2019, 39(5): 213-218.
LIN G, WANG B, PENG H, et al. Improved Faster-RCNN based multi-target detection and localization of transmission line inspection images[J]. Electric Power Automation Equipment, 2019, 39(5): 213-218.
[17] 王卓, 王玉静, 王庆岩, 等. 基于协同深度学习的二阶段绝缘子故障检测方法[J]. 电工技术学报, 2021, 36(17): 3594-3604.
WANG Z, WANG Y J, WANG Q Y, et al. Two-stage insulator fault detection method based on collaborative deep learning[J]. Transactions of China Electrotechnical Society, 2021, 36(17): 3594-3604.
[18] 王建烨, 续欣莹, 谢刚, 等. 改进SSD模型的绝缘子自爆故障检测[J]. 现代电子技术, 2022, 45(14): 115-121.
WANG J Y, XU X Y, XIE G, et al. Insulators self-explosion fault detection based on improved SSD model[J]. Modern Electronics Technique, 2022, 45(14): 115-121.
[19] 罗潇, 於锋, 彭勇, 等. 基于深度学习的无人机电网巡检缺陷检测研究[J]. 电力系统保护与控制, 2022, 50(10): 132-139.
LUO X, YU F, PENG Y, et al. UAV power grid inspection defect detection based on deep learning[J]. Power System Protection and Control, 2022, 50(10): 132-139.
[20] 王道累, 张世恒, 袁斌霞, 等. 基于改进YOLOv5的轻量化玻璃绝缘子自爆缺陷检测研究[J]. 高电压技术, 2023, 49(10): 4382-4390.
WANG D L, ZHANG S H, YUAN B X, et al. Research on self-explosion defect detection of lightweight glass insulators based on improved YOLOv5[J]. High Voltage Engineering, 2023, 49(10): 4382-4390.
[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[J]. arXiv:2207.02696, 2022.
[22] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]//Proceedings of the 15th European Conference on Computer Vision, 2018: 3-19.
[23] QYAN Y, ZHANG D, ZHANG L, et al. Centralized feature pyramid for object detection[J]. arXiv:2210.02093, 2022.