(英)空中输电线路巡检的LiDAR多通道光谱图像异常识别技术
Received:January 12, 2017  Revised:April 17, 2017  点此下载全文
引用本文:
Hits: 350
Download times: 64
Author NameAffiliationE-mail
REN Tian-yu Changchun University of Science and Technolog turbo@126.com 
DUANMU Qing-Duo Changchun University of Science and Technolog duanmu@cust.edu.cn 
wu bo qi Changchun University of Science and Technology  
jiang hui lin Changchun University of Science and Technology  
xu jin kai Changchun University of Science and Technology  
基金项目:Supported by Jilin Key Science and Technology Research Fund(130202GX010012794); National Natural Science Foundation of China (61077024,11474037); Specialized Research Fund for the Doctoral Program of Higher EducationofChina (20112216120005 );
中文摘要:光谱分析对于空中输电线路中的异常特征识别与机理研究有重要意义,本文根据长期以来的大量实验证明,多通道光谱分析法对输电线路绝缘子异常的多波段参数提取、模式识别、波形转换、异常目标识别具有极强的应用价值。为提高空中输电线路巡检的工作效率以及迅速、精确的排查异常故障,本文提出一体化机载LiDAR多通道光谱图像异常识别系统。以机载激光雷达(LiDAR)技术为基础,联合GPS、ISN、LiDAR、激光测距机等器件,构建目标超POS信息;通过计算最小视场分辨率、最小像元数、焦距等参数选定IGV-B1630C相机作为多光谱光学相机;将POS信息采集系统与多光谱相机组合成LiDAR多通道光谱图像异常识别系统。采用多通道匹配融合方法将紫外、红外、彩色三种图片进行融合,以Hough变换为基础,运用同族容器归纳法处理图片与超POS信息,从而确定输电线路的异常、疑似异常故障点。重点通过Hough变换、免疫遗传Snake模型算法、最小二乘方法拟合、解析椭圆形貌,解决绝缘子的异常识别问题。通过大量的工程实验表明,LiDAR多通道光谱图像异常识别系统识别效率高于预期值60%,大于人工识别值30%,优于通用航空直升机/无人机25%,是一种高效率的智能电网巡线排查手段。
中文关键词:输电线路巡检  激光点云雷达  多光谱图像异常识别  超POS信息
 
LiDAR multichannel spectral abnormal image recognition technology for inspecting aerial power transmission lines
Abstract:Abstract Spectral analysis is significant to research on the abnormal recognition characteristics and mechanism of transmission lines. Many past experiments have shown that the multichannel spectral analysis method strongly influences the abnormal insulation of transmission lines during the extraction of multi-waveband parameters, pattern recognition, waveform conversion, and abnormal target recognition. The aforementioned method aims to rapidly and accurately improve operation efficiency of the inspection process of aerial power transmission lines and the screening of abnormal faults. The position and orientation system (POS) information of a target can be obtained through airborne laser radar (LiDAR) technology combined with the Global Positioning System, an inertial navigation system, and a laser range finder. The IGV-B2630C camera is used as a multispectral camera to calculate the minimum field of view resolution, minimum pixel number, focal length, and other parameters. The LiDAR multichannel spectrum image recognition system is composed of the POS information acquisition system and the multispectral camera. The multichannel matching fusion method can produce ultraviolet, infrared, and color pictures. The image and super POS information can be processed by using the kin container induction method according to the Hough transform method. Finally, the abnormal and suspected abnormal points of a transmission line can be determined. The elliptical shape can be fitted and parsed using the Hough transform method, the immune genetic snake model algorithm, and the least squares method, which can rapidly solve anomaly recognition problem in the insulator. The average failure detection resolution of LiDAR multi-channel spectral image anomaly recognition system is 82.4%, and it is higher than the average for copter and manual detection of 24.05%. The proposed system is a highly efficient smart grid patrol screening method.
keywords:High-voltage line inspection, LiDAR  Multispectral image anomaly recognition, Super POS information
  View/Add Comment  Download reader

Copyright:《Journal of Infrared And Millimeter Waves》