基于改进YOLOX的继电保护压板状态校核方法*

doi:10.11985/2024.01.032

电气工程学报 ›› 2024, Vol. 19 ›› Issue (1): 299-307.doi: 10.11985/2024.01.032

扫码分享

基于改进YOLOX的继电保护压板状态校核方法^*

方国权¹(), 赵俊¹(), 陈昊¹(), 吴司颖¹(), 丁涛²(), 施涛³()

1.国网江苏省电力有限公司超高压分公司南京 211102
2.西安交通大学电气工程学院西安 710049
3.南京邮电大学先进技术研究院南京 210023

收稿日期:2023-04-15 修回日期:2023-10-31 出版日期:2024-03-25 发布日期:2024-04-25
通讯作者: 方国权，男，1994年生，硕士。主要从事继电保护及自动化相关工作。E-mail：220172691@seu.edu.cn
作者简介:赵俊，男，1989年生，硕士，工程师。主要从事继电保护及自动化相关工作。E-mail：zhaojun1100zj@163.com;
陈昊，男，1980年生，博士，研究员级高级工程师。主要从事变电二次运检工作。E-mail：pingfengma@163.com;
吴司颖，女，1995年生，硕士。主要从事继电保护及自动化相关工作。E-mail：1623231099@qq.com;
丁涛，男，1986年生，博士，教授。主要从事电力系统经济运行、电力市场、能源经济领域研究。E-mail：tding15@mail.xjtu.edu.cn
施涛，男，1982年生，博士，高级工程师。主要从事新能源与储能并网运行控制技术领域研究。E-mail：shitao@njupt.edu.cn
基金资助:
国家自然科学基金资助项目(62192753)

Verification Method of Relay Protection Platen State Based on the Improved YOLOX

FANG Guoquan¹(), ZHAO Jun¹(), CHEN Hao¹(), WU Siying¹(), DING Tao²(), SHI Tao³()

1. EHV Voltage Branch Company, State Grid Jiangsu Electric Power Co., Ltd., Nanjing 211102
2. School of Electrical Engineering, Xi’an Jiaotong University, Xi’an 710049
3. Institude of Advanced Technology, Nanjing University of Posts and Telecommunications, Nanjing 210023

Received:2023-04-15 Revised:2023-10-31 Online:2024-03-25 Published:2024-04-25

摘要/Abstract

摘要：

继电保护压板状态校核是变电站巡检的重要业务，然而现有校核方法多为人工读取和复诵核对，存在费力耗时以及安全风险的不足。为此，提出一种基于改进YOLOX的智能化校核方法。首先，该方法采用以注意力机制改进主干特征提取模块的YOLOX模型，实现压板状态识别和压板标签定位。其次，针对标签边界框标定的图像区域进行倾斜角矫正，并基于光学字符引擎和压板双重名称专业语料库实现文本识别，获得压板状态和双重名称的映射关系。最后，采用实际变电站的压板图像进行试验验证。结果表明，对于压板状态识别，所提方法具有更高的识别精度和鲁棒性，并且对于文本识别、倾斜角矫正和专业语料库的构建，可以显著提高压板双重名称的识别准确率。所提方法为巡检机器人实现智能化压板状态校核提供了一种新思路。

关键词: 智能运检, 继电保护, 压板状态, 文本识别, 注意力机制, 透视变换

Abstract:

Verifying the state of relay protection plate is an important part of the substation inspection, but the existing verification methods are mostly manual reading and repetition checking, which is inefficiency and error-prone. Therefore, an intelligent verification method of protection plate state based on the improved YOLOX is proposed. Firstly, the method recognizes the platen state and locates platen label position based on YOLOX model which is improved by attention mechanism in the backbone feature extraction network. Secondly, the angle of the image area calibrated by the bounding box is corrected, and then text recognition is realized based on the optical character engine and professional corpus of the plate dual name in order to obtain the mapping relationship between the state and the dual name. Finally, images collected in actual substation is used for experiment. The results show that, for the recognition of the platen state, this proposed method has higher recognition accuracy and robustness, and for text recognition, skew angle correction and the construction of professional corpus can significantly improve the recognition accuracy of platen dual names. This proposed method provides a new idea for the smart inspection robot to realize the intelligent platen state check.

Key words: Intelligent maintenance, relay protection, plate status, text recognition, attention mechanism, perspective transform

中图分类号:

TM774

方国权, 赵俊, 陈昊, 吴司颖, 丁涛, 施涛. 基于改进YOLOX的继电保护压板状态校核方法^*[J]. 电气工程学报, 2024, 19(1): 299-307.

FANG Guoquan, ZHAO Jun, CHEN Hao, WU Siying, DING Tao, SHI Tao. Verification Method of Relay Protection Platen State Based on the Improved YOLOX[J]. Journal of Electrical Engineering, 2024, 19(1): 299-307.

图/表 13

图1

图2

图3

图4

图5

图6

表1

图7

图8

表2

表3

图9

表4

参考文献 22

[1]	李金, 张喜铭, 谢型浪, 等. 变电站安全运维装置技术方案[J]. 电气工程学报, 2020, 15(4):114-120.
	LI Jin, ZHANG Ximing, XIE Xinglang, et al. Technical scheme of substation safety operation and maintenance device[J]. Journal of Electrical Engineering, 2020, 15(4):114-120.
[2]	周立辉, 张永生, 孙勇, 等. 智能变电站巡检机器人研制及应用[J]. 电力系统自动化, 2011, 35(19):85-88.
	ZHOU Lihui, ZHANG Yongsheng, SUN Yong, et al. Development and application of equipment inspection robot for smart substations[J]. Automation of Electric Power Systems, 2011, 35(19):85-88.
[3]	谢林枫, 蒋超, 孙秋芹, 等. 基于AMC算法的变电站巡检机器人地图创建与定位[J]. 电力工程技术, 2019, 38(5):16-23.
	XIE Linfeng, JIANG Chao, SUN Qiuqin, et al. The global map’s creating and positioning of substation inspection robot based on adaptive Monte Carlo particle filter algorithm[J]. Electrical Power Engineering Technology, 2019, 38(5):16-23.
[4]	袁朝晖, 付文龙, 李佰霖, 等. 基于多策略分割融合与形态特征辨识的变电站保护压板状态识别[J]. 电力系统保护与控制, 2022, 50(1):98-106.
	YUAN Zhaohui, FU Wenlong, LI Bailin, et al. Protection platen status recognition for a smart substation based on multi-strategy segmentation and fusion and morphological feature identification[J]. Power System Protection and Control, 2022, 50(1):98-106.
[5]	付文龙, 谭佳文, 吴喜春, 等. 基于图像处理与形态特征分析的智能变电站保护压板状态识别[J]. 电力自动化设备, 2019, 39(7):203-207.
	FU Wenlong, TAN Jiawen, WU Xichun, et al. Protection platen status recognition based on image processing and morphological feature analysis for smart substation[J]. Electric Power Automation Equipment, 2019, 39(7):203-207.
[6]	梁辰, 孙建文, 王兰玉, 等. 应用聚类和证据理论实现变电站保护压板状态校核[J]. 电网技术, 2020, 44(6):2343-2349.
	LIANG Chen, SUN Jianwen, WANG Lanyu, et al. Application of clustering and evidence theory to checking the status of substation protection platen[J]. Power System Technology, 2020, 44(6):2343-2349.
[7]	王伟, 张彦龙, 翟登辉, 等. 基于OpenCV+SSD深度学习模型的变电站压板状态智能识别[J]. 电测与仪表, 2022, 59(1):106-112.
	WANG Wei, ZHANG Yanlong, ZHAI Denghui, et al. Intelligent identification of substation platen state based on OpenCV+SSD deep learning model[J]. Electrical Measurement & Instrumentation, 2022, 59(1):106-112.
[8]	周克, 杨倩文, 王耀艺, 等. 一种改进的压板状态识别SSD算法[J]. 电测与仪表, 2021, 58(1):69-76.
	ZHOU Ke, YANG Qianwen, WANG Yaoyi, et al. An improved SSD algorithm for pressure platestatus recognition[J]. Electrical Measurement & Instrumentation, 2021, 58(1):69-76.
[9]	汪洋, 黎恒烜, 鄂士平, 等. 基于深度学习的变电站硬压板状态检测与识别算法[J]. 沈阳工业大学学报, 2020, 42(6):676-680. doi: 10.7688/j.issn.1000-1646.2020.06.11
	WANG Yang, LI Hengxuan, E Shiping, et al. State detection and recognition algorithm for hard platens of substation based on deep learning[J]. Journal of Shenyang University of Technology, 2020, 42(6):676-680. doi: 10.7688/j.issn.1000-1646.2020.06.11
[10]	ULLAH M B. CPU based YOLO:A real time object detection algorithm[C]// 2020 IEEE Region 10 Symposium (TENSYMP), June 5-7,2020,Dhaka,Bangladesh,2020:552-555.
[11]	LIU Bingqian, HUANG Jianye, LIN Shuang, et al. Improved YOLOX-S abnormal condition detection for power transmission line corridors[C]// IEEE 3rd International Conference on Power Data Science(ICPDS), December 26-26,2021,Harbin,China,2021:13-16.
[12]	LI Youyou, HE Binbin, MELGANI F, et al. Point-based weakly supervised learning for object detection in high spatial resolution remote sensing images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14(2):5361-5371. doi: 10.1109/JSTARS.2021.3076072
[13]	WANG C Y, LIAO H Y, WU Y H, et al. CSPNet:A new backbone that can enhance learning capability of CNN[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops(CVPRW), June 14-19,2020,Seattle,WA,USA,2020:1571-1580.
[14]	LIN T Y, DOLLAR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition, July 21-26,2017,Honolulu,HI,USA,2017:936-944.
[15]	蔡逢煌, 张岳鑫, 黄捷. 基于YOLOv3与注意力机制的桥梁表面裂痕检测算法[J]. 模式识别与人工智能, 2020, 33(10):926-933. doi: 10.16451/j.cnki.issn1003-6059.202010007
	CAI Fenghuang, ZHANG Yuexin, HUANG Jie. Bridge surface crack detection algorithm based on YOLOv3 and attention mechanism[J]. Pattern Recognition and Artificial Intelligence, 2020, 33(10):926-933. doi: 10.16451/j.cnki.issn1003-6059.202010007
[16]	GAO Zishu, YANG Guodong, LI En, et al. Novel feature fusion module-based detector for small insulator defect detection[J]. IEEE Sensors Journal, 2021, 21(15):16807-16814. doi: 10.1109/JSEN.2021.3073422
[17]	羊冰清, 张周燕, 朱小良. 基于数字图像处理的液位测量系统的研究与实现[J]. 电力工程技术, 2018, 37(4):57-64.
	YANG Bingqing, ZHANG Zhouyan, ZHU Xiaoliang. Study and implement on the measurement system of liquid level based on digital image processing[J]. Electrical Power Engineering Technology, 2018, 37(4):57-64.
[18]	操昊鹏, 曾卫明, 石玉虎, 等. 基于Hough变换和总体最小二乘法的电力线检测[J]. 计算机技术与发展, 2018, 28(10):164-167.
	CAO Haopeng, ZENG Weiming, SHI Yuhu, et al. Power line detection based on Hough transform and total least squares method[J]. Computer Technology and Development, 2018, 28(10):164-167.
[19]	李思妍, 台升, 张宇航, 等. 基于轻量化YOLOv3和Tesseract OCR的电力设备标志牌识别技术[J]. 智慧电力, 2021, 49(7):79-85,108.
	LI Siyan, TAI Sheng, ZHANG Yuhang, et al. Electrical sign recognition technology based on simplified YOLOv3 and Tesseract OCR[J]. Smart Power, 2021, 49(7):79-85,108.
[20]	陈昊, 姚凯, 张海华, 等. 变电站设备缺陷图像识别的实用化评价方法[J]. 湖北电力, 2021, 45(5):52-57.
	CHEN Hao, YAO Kai, ZHANG Haihua, et al. A practical evaluation method for image recognition algorithm of device defects in substations[J]. Hubei Electric Power, 2021, 45(5):52-57.
[21]	WANG Aili, WANG Minhui, JIANG Kaiyuan, et al. A novel lidar data classification algorithm combined densenet with STN[C]// 2019 IEEE International Geoscience and Remote Sensing Symposium,July 28-August 2, 2019,Yokohama,Japan,2019:2483-2486.
[22]	刘学平, 李玙乾, 刘励, 等. 嵌入SENet结构的改进YOLOV3目标识别算法[J]. 计算机工程, 2019, 45(11):243-248. doi: 10.19678/j.issn.1000-3428.0052861
	LIU Xueping, LI Yuqian, LIU Li, et al. Improved YOLOV3 target recognition algorithm with embedded SENet structure[J]. Computer Engineering, 2019, 45(11):243-248. doi: 10.19678/j.issn.1000-3428.0052861

训练参数	冻结阶段	解冻阶段
训练轮次	200	200
训练批量大小	24	8
学习率	0.001	0.000 1

注意力机制模块	精确率(%)	召回率(%)	平均准确度均值(%)
STN^[21]	81.2	93.3	89.2
SENet^[22]	83.7	94.1	91.6
CBAM	86.1	97.6	94.1

模型	平均准确度均值(%)	参数大小/ MB	每秒检测数FPS
Faster-RCNN	95.3	135	21.3
YOLOv3	91.1	58.3	60.8
YOLOX	92.8	31.5	75.5
改进YOLOX	98.1	34	73.1

改进方式		非中文字符Acc(%)
自带语料库	58.4	47.2
训练后专业语料库	90.2	91.8
专业语料库+倾斜角矫正	96.6	96.0

基于改进YOLOX的继电保护压板状态校核方法^*

Verification Method of Relay Protection Platen State Based on the Improved YOLOX

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 13

参考文献 22

相关文章 11

编辑推荐

Metrics

本文评价

[1]	解骞, 徐浩岚, 王彤, 赵发寿, 张刚, 党建. 基于自主认知深度时间聚类表示的隔离开关故障诊断方法^*[J]. 电气工程学报, 2024, 19(1): 281-289.
[2]	任爽, 杨凯, 商继财, 祁继明, 魏翔宇, 蔡永根. 基于CNN-BiGRU-Attention的短期电力负荷预测[J]. 电气工程学报, 2024, 19(1): 344-350.
[3]	刘昕明, 吉建光, 李玮, 石光磁. 基于双模态分解的发电站母线短期负荷预测^*[J]. 电气工程学报, 2024, 19(1): 124-132.
[4]	李志兵, 肖健梅, 王锡淮. 基于多粒度NRS和改进Bi-LSTM的电力系统暂态稳定评估^*[J]. 电气工程学报, 2023, 18(3): 232-241.
[5]	余晓霞, 汤宝平, 王伟影, 吴宣勇, 李彪. 复杂工况条件下多头注意力双向长短时记忆网络的风电机组缺失数据修复方法研究[J]. 机械工程学报, 2023, 59(14): 1-9.
[6]	卞文彬, 邓艾东, 刘东川, 赵敏, 刘洋, 李晶. 基于改进深度残差收缩网络的风电机组滚动轴承故障诊断方法[J]. 机械工程学报, 2023, 59(12): 202-214.
[7]	吴彤, 李冰锋, 费树岷, 连东辉. 基于改进FCOS算法的架空输电线路防振锤检测^*[J]. 电气工程学报, 2023, 18(1): 143-152.
[8]	鲁南, 欧阳权, 黄俍卉, 王志胜. 基于注意力机制和多任务LSTM的锂电池容量预测方法^*[J]. 电气工程学报, 2022, 17(4): 41-50.
[9]	杨炜, 黄立红, 屈晓磊. 融合双注意力机制EfficientNet的沥青路面状态分类方法[J]. 机械工程学报, 2022, 58(24): 211-222.
[10]	刘永刚, 于丰宁, 章新杰, 陈峥, 秦大同. 基于激光点云与图像融合的3D目标检测研究[J]. 机械工程学报, 2022, 58(24): 289-299.
[11]	董绍江, 裴雪武, 吴文亮, 汤宝平, 赵兴新. 基于多层降噪技术及改进卷积神经网络的滚动轴承故障诊断方法[J]. 机械工程学报, 2021, 57(1): 148-156.