基于无人机图像识别技术的输电线路缺陷检测
|
摘要
针对无人机巡检输电线路过程中,存在人工识别缺陷图像工作量大、效率低的问题,利用深度卷积神经网络算法开发无人机巡检数据智能管理平台,满足无人机飞行作业管理应用的需要。利用无人机巡视输电线的便利性,探究了基于无人机图像识别的无人机立体智能巡检应用平台的系统结构,采用深度卷积神经网络算法对无人机日常巡检产生的海量图像或视频数据进行预处理识别。同时,对图像缺陷的位置进行标注并进行种类分类,将识别结果反馈,形成标准的缺陷报告。对于20种物体的分类任务,精度达到了85%左右;对于常见的输电设备缺陷达到了75%的精度。测试结果表明,该算法能准确识别出输电线路缺陷图像,提高了无人机巡视缺陷的效率,且降低用人成本。
In order to solve the problem of large workload and low efficiency of manual identification of defective images in the process of Unmanned Aerial Vehicle(UAV)patrol inspection of transmission lines,an intelligent management platform of uav patrol data is developed using deep convolution neural network algorithm to meet the needs of uav flight operation management application. Utilizing the convenience of UAVs to patrol the transmission lines,the system structure of the UAV stereo intelligent inspection application platform based on UAV image recognition is explored. The deep convolutional neural network algorithm is used to generate a large amount of daily inspections of UAVs. The image or video data is preprocessed and classified,and the position of the image defect is marked and classified,and the recognition result is fed back to form a standard defect report. For the classification task of 20 kinds of objects,the precision reaches about 85%;for the common transmission equipment defects,the accuracy is 75%.The test results show that the algorithm can accurately identify the defect image of the transmission line,which improves the efficiency of the UAV inspection defect and reduce the cost of the user.
In order to solve the problem of large workload and low efficiency of manual identification of defective images in the process of Unmanned Aerial Vehicle(UAV)patrol inspection of transmission lines,an intelligent management platform of uav patrol data is developed using deep convolution neural network algorithm to meet the needs of uav flight operation management application. Utilizing the convenience of UAVs to patrol the transmission lines,the system structure of the UAV stereo intelligent inspection application platform based on UAV image recognition is explored. The deep convolutional neural network algorithm is used to generate a large amount of daily inspections of UAVs. The image or video data is preprocessed and classified,and the position of the image defect is marked and classified,and the recognition result is fed back to form a standard defect report. For the classification task of 20 kinds of objects,the precision reaches about 85%;for the common transmission equipment defects,the accuracy is 75%.The test results show that the algorithm can accurately identify the defect image of the transmission line,which improves the efficiency of the UAV inspection defect and reduce the cost of the user.
引文
[1]张浩.输电线路缺陷在线监控系统设计与实现[D].成都:电子科技大学,2013.
[2]韩冰,尚方.面向无人机输电线路巡检的电力杆塔检测框架模型[J].浙江电力,2016,35(4):6-11.
[3]郭昕阳,李卫国,郑建刚,等.一种轻型电动无人机在输电线路巡视中的应用技术[J].华北电力技术,2013(4):61-64.
[4]韦舒天,李龙,岳灵平,等.输电通道人机协同巡检方式的探索[J].浙江电力,2016,35(3):10-13.
[5]杨蔚,周辉,杨生兰,等.输电线路巡检绝缘子跟踪与测距技术[J].电子科技,2017,30(5):124-127.
[6] Larrauri JI,Sorrosal G,Gonzalez M.Automatic system for overhead power line inspection using an unmanned aerial vehicle-RELIFO project[C].Boston:2013 International Conference on Unmanned Aircraft Systems(ICUAS),IEEE,2013.
[7] Dong G,Chen X,Wang B,et al.Inspecting transmission lines with an unmanned fixed-wings aircraft[C]. Wuhan:2012 2nd International Conference on Applied Robotics for the Power Industry(CARPI),IEEE,2012.
[8]汤明文,戴礼豪,林朝辉,等.无人机在电力线路巡视中的应用[J].中国电力,2013(3):35-38.
[9]姚佳,唐国红,谌玲.基于压缩感知的输电线路状态监视图像传输[J].电子科技,2014,27(4):159-162.
[10]王淼,杜毅,张忠瑞.无人机辅助巡视及绝缘子缺陷图像识别研究[J].电子测量与仪器学报,2015,29(12):1862-1868.
[11]殷栢辉,钟凯宇,张学习.基于无人机图像的输电线断股缺陷实时检测[J].自动化与信息工程,2016,37(4):172-176.
[12]王亚萍,韩军,陈舫明,等.可见光图像中的高压线缺陷自动诊断方法[J].计算机工程与应用,2011,47(12):180-184.
[13]程嘉晖,基于深度卷积神经网络飞行器图像识别算法研究[D].杭州:浙江大学,2017.
[14]张志远.基于深度学习的光学遥感图像飞机检测[D].厦门:厦门大学,2016.
[15]赵振兵,王乐.一种航拍绝缘子串图像自动定位方法[J].仪器仪表学报,2014,12(3):558-565.
[16]赵建青,姚瑶,邱玩辉,等.基于输电线路在线巡视系统的智能预警系统研究[J].电力系统保护与控制,2013,41(23):49-54.
[2]韩冰,尚方.面向无人机输电线路巡检的电力杆塔检测框架模型[J].浙江电力,2016,35(4):6-11.
[3]郭昕阳,李卫国,郑建刚,等.一种轻型电动无人机在输电线路巡视中的应用技术[J].华北电力技术,2013(4):61-64.
[4]韦舒天,李龙,岳灵平,等.输电通道人机协同巡检方式的探索[J].浙江电力,2016,35(3):10-13.
[5]杨蔚,周辉,杨生兰,等.输电线路巡检绝缘子跟踪与测距技术[J].电子科技,2017,30(5):124-127.
[6] Larrauri JI,Sorrosal G,Gonzalez M.Automatic system for overhead power line inspection using an unmanned aerial vehicle-RELIFO project[C].Boston:2013 International Conference on Unmanned Aircraft Systems(ICUAS),IEEE,2013.
[7] Dong G,Chen X,Wang B,et al.Inspecting transmission lines with an unmanned fixed-wings aircraft[C]. Wuhan:2012 2nd International Conference on Applied Robotics for the Power Industry(CARPI),IEEE,2012.
[8]汤明文,戴礼豪,林朝辉,等.无人机在电力线路巡视中的应用[J].中国电力,2013(3):35-38.
[9]姚佳,唐国红,谌玲.基于压缩感知的输电线路状态监视图像传输[J].电子科技,2014,27(4):159-162.
[10]王淼,杜毅,张忠瑞.无人机辅助巡视及绝缘子缺陷图像识别研究[J].电子测量与仪器学报,2015,29(12):1862-1868.
[11]殷栢辉,钟凯宇,张学习.基于无人机图像的输电线断股缺陷实时检测[J].自动化与信息工程,2016,37(4):172-176.
[12]王亚萍,韩军,陈舫明,等.可见光图像中的高压线缺陷自动诊断方法[J].计算机工程与应用,2011,47(12):180-184.
[13]程嘉晖,基于深度卷积神经网络飞行器图像识别算法研究[D].杭州:浙江大学,2017.
[14]张志远.基于深度学习的光学遥感图像飞机检测[D].厦门:厦门大学,2016.
[15]赵振兵,王乐.一种航拍绝缘子串图像自动定位方法[J].仪器仪表学报,2014,12(3):558-565.
[16]赵建青,姚瑶,邱玩辉,等.基于输电线路在线巡视系统的智能预警系统研究[J].电力系统保护与控制,2013,41(23):49-54.