本文在人工智能的基础上利用深度学习结合红外光热成像技术对无人机巡检技术进行研究和优化。首先对红外光热成像技术在电力系统中的应用和无人机巡检技术进行了介绍;然后提出了基于卷积神经网络的目标检测算法,对卷积神经网络的计算原理以及目标检测算法进行说明,并设计了基于深度视觉系统的MobileNetV1YOLOv3网络预测模型。通过实测分析发现,基于深度学习的轻量级的目标检测网络预测模型定位误差在X、Y、Z三个维度上均低于GPS系统,最小误差仅为0.06 m。
Abstract
This article uses deep learning and infrared thermal imaging technology to research and optimize the drone inspection technology on the basis of artificial intelligence.Firstly, the application of infrared thermal imaging technology in the power system and the intelligent UAV inspection technology are introduced.Secondly, a target detection algorithm based on convolutional neural network is proposed, the calculation principle of convolutional neural network and target detection algorithm are explained, and the MobileNetV1YOLOv3 network prediction model based on deep vision system is designed on this basis.Through actual measurement and analysis, the positioning error of the target detection network prediction model is lower than that of the GPS system in the three dimensions of X, Y and Z, and the minimum error is only 0.06 m.
关键词
深度学习 /
红外光热成像 /
无人机巡检 /
输电线检测 /
MobileNetV1YOLOv3
Key words
deep learning /
infrared thermal imaging /
UAV inspection /
transmission line inspection /
MobileNetV1YOLOv3
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 徐敏, 罗劲, 刘华.基于Faster RCNN级联的输电塔螺栓缺失检测[J].广东电力, 2021, 34(6):6571.
[2] Chaudhry Muhammad, Hasan Md Mahmudul, Erkinbaev Chyngyz, et al.Bison muscle discrimination and color stability prediction using nearinfrared hyperspectral imaging[J].Biosystems Engineering, 2021, 209(5):32.
[3] 李国强, 彭炽刚, 汪勇, 等.基于深度学习的杆塔三维姿态实时估计[J].电子技术应用, 2021, 47(6):8791, 95.
[4] Polewski Przemyslaw, Shelton Jacquelyn, Yao Wei, et al.Instance segmentation of fallen trees in aerial color infrared imagery using active multicontour evolution with fully convolutional networkbased intensity priors[J].ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 178(5):64.
[5] 张涛, 高绅, 陈万培, 等.基于深度学习的间隔棒故障识别方法[J].无线电工程, 2021, 51(6):502506.
[6] 丁志敏, 邢晓敏, 董行, 等.基于深度学习的输电线挂接地线状态目标检测[J].电子测量技术, 2021, 44(3):132137.
[7] 苏素平, 李虹, 孙志毅, 等.无人机图像的输电线断股检测方法研究[J].太原科技大学学报, 2021, 42(1):3236.
[8] Jiménez Balado Joan, Eich Teal S.GABAergic dysfunction, neural network hyperactivity and memory impairments in human aging and Alzheimer’s disease[J].Seminars in Cell&Developmental Biology, 2021, 116(5):72.
[9] Regivaldo Carvalho, Richardson Nascimento, Uzeda Garcia, et al.A UAVBased Framework for SemiAutomated Thermographic Inspection of Belt Conveyors in the Mining Industry[J].Sensors, 2020, 20(8):53.
[10] 王红星, 翟学锋, 陈玉权.基于改进Cascade RCNN的两阶段销钉缺陷检测模型[J].科学技术与工程, 2021, 21(15):63736379.
[11] 应斌, 唐斌, 潘俊杰, 等.基于YOLOv3的电力杆塔检测算法研究[J].浙江电力, 2021, 40(5):5359.
[12] 韦立富, 沈佩琦, 韩磊, 等.5G无人机全自主巡检系统在500kV特大跨越上的应用[J].电工技术, 2021, 4(9):7376.
[13] Jae Kang Lee, Jung Ok Kim, Seong Jin Park.A study on the UAV imagebased efficiency improvement of bridge maintenance and inspection[J].Journal of Intelligent & Fuzzy Systems, 2018, 36(2):129.
[14] 陈同凡, 刘云鹏, 裴少通.基于改进YOLOv3的电力设备红外诊断方法[J].广东电力, 2021, 34(6):2129.
[15] 崔平.红外成像在铜冶炼圆盘浇铸机上的智能化应用[J].有色冶金设计与研究, 2021, 42(3):3638.
[16] 戚海洋, 李守恒.电梯电气系统检测中的红外热成像技术应用[J].电子技术, 2021, 50(6):1617.
[17] 杨东, 陈瑞斌, 仝海龙.基于红外检测技术的电力系统变压器套管绝缘缺陷故障检测[J].自动化与仪表, 2021, 36(5):5761.
[18] 赵晓鹏, 郭威.无人机图像的输电线异物检测方法研究[J].太原科技大学学报, 2021, 42(2):104108.
[19] 李海峰, 武剑灵, 李立学.YOLOv3ad在输电线路监控中提高小目标物体检测准确度的研究[J].广东电力, 2021, 34(2):92100.
PDF(1015 KB)
