水电站智能巡检机器人技术的应用

doi:10.16180/j.cnki.issn1007-7820.2023.12.014

电子科技 ›› 2023, Vol. 36 ›› Issue (12): 99-102.doi: 10.16180/j.cnki.issn1007-7820.2023.12.014

• • 上一篇

水电站智能巡检机器人技术的应用

沈浩,赵毅锋,李晓

华东天荒坪抽水蓄能有限责任公司,浙江湖州 313302

收稿日期:2022-04-06 出版日期:2023-12-15 发布日期:2023-12-05
作者简介:沈浩(1993-),男,助理工程师。研究方向:抽水蓄能电站水工运维与管理。
基金资助:
国网新能源控股有限公司科研项目(SGXYTP00JHJS1800108)

Application of Intelligent Inspection Robot Technology for Hydropower Station

SHEN Hao,ZHAO Yifeng,LI Xiao

East China Tianhuangping Pumped Storage Co., Ltd., Huzhou 313302, China

Received:2022-04-06 Online:2023-12-15 Published:2023-12-05
Supported by:
Research Project of State Grid New Energy Holdings Co., Ltd.(SGXYTP00JHJS1800108)

摘要/Abstract

摘要：

针对传统基于人工抽水蓄能站裂缝与渗水检测成本过高、检测精度难以保证等问题,文中设计了一套基于机器视觉的巡检机器人系统。构建综合交叉熵与Dice代价函数的卷积神经网络,建立基于总像素准确率、交并比与F1-score的评价函数,确保准确检测常见裂缝。为了验证机器人巡检系统的有效性,文中检验了卷积神经网络,并与常见的计算机视觉方法与人工检测方法进行性能对比。对比结果证明,文中构建的神经网络在检测准确性与检测效率方面均有明显进步。

关键词: 裂缝检测, CNN, 代价函数, 计算机视觉, 巡检机器人, 水利系统

Abstract:

In view of the high cost of crack and seepage detection in the traditional artificial pumped storage station and the difficulty of ensuring the detection accuracy, a set of inspection robot system based on machine vision is designed in this study. A convolutional neural network which combines cross entropy and dice cost function is constructed, and an evaluation function based on total pixel accuracy, cross parallel ratio and F1-score is established to ensure the accurate detection of common cracks. In order to verify the effectiveness of the designed robot inspection system, convolutional neural network is tested in this study, and its performance is compared with common computer vision methods and manual detection methods. The comparison results show that the neural network constructed in this study has obvious progress in detection accuracy and detection efficiency.

Key words: crack detection, CNN, cost function, computer vision, inspection robot, hydraulic system

中图分类号:

TP39

沈浩,赵毅锋,李晓. 水电站智能巡检机器人技术的应用[J]. 电子科技, 2023, 36(12): 99-102.

SHEN Hao,ZHAO Yifeng,LI Xiao. Application of Intelligent Inspection Robot Technology for Hydropower Station[J]. Electronic Science and Technology, 2023, 36(12): 99-102.

图/表 6

图1

图2

图3

图4

表1

表2

参考文献 10

[1]	牛连强, 陈向震, 张胜男, 等. 深度连续卷积神经网络模型构建与性能分析[J]. 沈阳工业大学学报, 2016, 38(6):662-666. doi: 10.7688/j.issn.1000-1646.2016.06.12
	Niu Lianqiang, Chen Xiangzhen, Zhang Shengnan, et al. Construction and performance analysis of deep continuous convolution neural network model[J]. Journal of Shenyang University of Technology, 2016, 38(6):662-666. doi: 10.7688/j.issn.1000-1646.2016.06.12
[2]	王培雷. 基于残差网络的多标签图像检索方法研究[D]. 桂林: 桂林电子科技大学, 2019:35-60.
	Wang Peilei. Research on multi tag image retrieval method based on residual network[D]. Guilin:Guilin University of Electronic Science and Technology, 2019:35-60.
[3]	王伟栋, 李菲菲, 谢林, 等. 基于深度神经网络和哈希算法的图像检索研究[J]. 电子科技, 2018, 31(10):48-52.
	Wang Weidong, Li Feifei, Xie Lin, et al. Research on image retrieval based on depth neural network and hash algorithm[J]. Electronic Science and Technology, 2018, 31(10):48-52.
[4]	郜丽鹏, 郑辉. 基于PReLUs-Softplus非线性激励函数的卷积神经网络[J]. 沈阳工业大学学报, 2018, 40(1):54-59. doi: 10.7688/j.issn.1000-1646.2018.01.10
	Gao Lipeng, Zheng Hui. Convolution neural network based on prelus softplus nonlinear excitation function[J]. Journal of Shenyang University of Technology, 2018, 40(1):54-59. doi: 10.7688/j.issn.1000-1646.2018.01.10
[5]	梁珊, 刘晓波. 基于交叉熵代价函数的DBN转子系统故障识别分析[J]. 南昌航空大学学报:自然科学版, 2018, 32(3):1-7.
	Liang Shan, Liu Xiaobo. Fault identification analysis of DBN rotor system based on cross entropy cost function[J]. Journal of Nanchang University of Aeronautics and Technology (Natural Science Edition), 2018, 32(3):1-7.
[6]	刘昊, 王冠华, 章强, 等. 3D脑肿瘤分割的Dice损失函数的优化[J]. 中国医疗设备, 2019, 34(5):20-23,31.
	Liu Hao, Wang Guanhua, Zhang Qiang, et al. Optimization of dice loss function in 3D brain tumor segmentation[J]. Chinese Medical Equipment, 2019, 34(5):20-23,31.
[7]	李海涛, 戴莉莉, 顾海燕, 等. 样本尺寸对遥感影像FCN训练模型的影响分析[J]. 测绘科学, 2019, 44(6):133-137.
	Li Haitao, Dai Lili, Gu Haiyan, et al. Analysis of the influence of sample size on FCN training model of remote sensing image[J]. Surveying and Mapping Science, 2019, 44(6):133-137.
[8]	盛恒, 黄铭, 杨晶晶. 基于FasterR-CNN和IoU优化的实验室人数统计与管理系统[J]. 计算机应用, 2019, 39(6):1669-1674. doi: 10.11772/j.issn.1001-9081.2018102182
	Sheng Heng, Huang Ming, Yang Jingjing. Laboratory population statistics and management system optimized by FasterR-CNN and IoU[J]. Computer Application, 2019, 39(6):1669-1674. doi: 10.11772/j.issn.1001-9081.2018102182
[9]	黄浩, 徐海华, 王羡慧, 等. 自动发音错误检测中基于最大化F1值准则的区分性特征补偿训练算法[J]. 电子学报, 2015, 43(7):1294-1299.
	Huang Hao, Xu Haihua, Wang Xianhui, et al. Training algorithm of differential feature compensation based on maximizing F1 value in automatic pronunciation error detection[J]. Journal of Electronics, 2015, 43(7):1294-1299.
[10]	黄浩, 王建明, 哈力旦·阿布都热依木, 等. 自动发音错误检测中基于F1值最大化的声学模型训练方法[J]. 声学学报, 2013, 38(6):751-758.
	Huang Hao, Wang Jianming, Khalidin Abdureyimu, et al. Acoustic model training method based on F1 value maximization in automatic pronunciation error detection[J]. Journal of Acoustics, 2013, 38(6):751-758.

	CNN	Scharr	Snake	图搜索
检测时间/ms	320	480	470	590
误检率/%	0.2	3.2	6.2	2.0
准确性/%	98	87	91	95

拍摄距离 /mm	人工测量宽度/mm	CNN测量宽度/mm	绝对误差 /mm	相对误差 /%
1 403	0.35	0.34	-0.01	-2.8
1 076	0.36	0.37	0.01	-2.7
2 012	0.78	0.77	-0.01	-1.2
2 858	0.89	0.89	0.00	0.0
2 299	0.78	0.80	0.02	2.6
2 954	0.67	0.70	0.03	4.4
1 076	0.56	0.54	-0.02	-3.6
1 298	0.78	0.76	-0.02	-2.6

[1]	史建柯,乔美英,李冰锋,赵岩. 基于注意力机制的水下遮挡目标检测算法[J]. 电子科技, 2023, 36(5): 62-70.
[2]	伞红军,王汪林,陈久朋,谢飞亚,徐洋洋,陈佳. 面向室内动态场景的VSLAM[J]. 电子科技, 2022, 35(4): 14-19.
[3]	陈鹏,刘子龙. 基于GAN-CNN的心律失常识别[J]. 电子科技, 2022, 35(3): 45-50.
[4]	袁子昂,倪伟,冉敬楠. 基于Winograd算法的可重构卷积神经网络加速器[J]. 电子科技, 2022, 35(12): 35-42.
[5]	张秋元,刘志全. 一种基于纹理特征分析的地铁隧道裂缝检测方案[J]. 电子科技, 2022, 35(11): 36-41.
[6]	陈文佑,章伟,史晓帆,宋芳. 一种改进ORB特征匹配的半稠密三维重建ORB-SLAM算法[J]. 电子科技, 2021, 34(12): 62-67.
[7]	张莹,刘子龙,万伟. 基于Faster R-CNN的无人机车辆目标检测[J]. 电子科技, 2021, 34(11): 11-20.
[8]	苏烨,李筠,杨海马,刘瑾,江声华. 基于Faster R-CNN的钢轨表面缺陷识别研究[J]. 电子科技, 2020, 33(9): 63-68.
[9]	黎阳,沈烨,刘敏,戴仁月,姜晓燕. 融合运动信息与表观信息的多目标跟踪算法[J]. 电子科技, 2020, 33(9): 21-25.
[10]	周文凯,韩芳,孔维健. 基于Faster-RCNN的极验点选式验证码识别[J]. 电子科技, 2019, 32(9): 42-46.
[11]	毛晨悦,吴鹏勇. 基于人工势场法的无人机路径规划避障算法[J]. 电子科技, 2019, 32(7): 65-70.
[12]	袁小平,王岗,王晔枫,汪喆远,孙辉. 基于改进卷积神经网络的交通标志识别方法[J]. 电子科技, 2019, 32(11): 28-32.
[13]	刘云波,黄华. 施工现场安全帽佩戴情况监控技术研究[J]. , 2015, 28(4): 69-.
[14]	赵日成. 基于车底阴影的车前障碍物检测[J]. , 2015, 28(3): 15-.
[15]	马立新,周小波,单宇. 基于自适应PCNN的紫外图像融合算法[J]. , 2015, 28(10): 1-.

水电站智能巡检机器人技术的应用

Application of Intelligent Inspection Robot Technology for Hydropower Station

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 10

相关文章 15

Metrics

本文评价

推荐阅读 0