基于光流场的输电导线舞动监测方法  

Transmission line galloping monitoring based on optical flow field

作　　者：黄新波[1,2] 陶保震[3] 冯玲[1] 

机构地区：[1]西安工程大学电子信息学院,陕西西安710048 [2]华南理工大学电力学院,广东广州510640 [3]国电南京自动化股份有限公司,江苏南京210003

出　　处：《电力自动化设备》2012年第7期99-103,共5页Electric Power Automation Equipment

基　　金：国家自然科学基金资助项目(51177115);国家重点基础研究发展计划项目(973项目)(2009CB724507-3);中国博士后科学基金特别资助项目(201003348);中国博士后基金资助项目(20090460759)~~

摘　　要：提出基于光流场的输电导线舞动监测方法,以输电导线舞动视频中截取的多帧连续图像为研究对象,利用金字塔光流法计算图像中各点的光流值,通过形态学滤波中的闭运算滤除孤立噪声点,设置面积阈值提取图像中舞动的导线;对图像进行多层亚像素级滤波处理,把图像中向量场以横向、纵向分为实值图,以横向、纵向估算光流像素点的位置并计算出光流速度;描绘出图像中各点的运动方向和位移,实现导线舞动的自动监测。该方法克服了视频监测单纯依靠工作人员观看导线运行状况视频的弊端。实例分析结果表明,该方法能够准确地检测出导线舞动的目标区域并计算出导线上各点的速度和舞动位移。The galloping monitoring of power transmission line is put forward based on optical flow field, which takes the multi-frame images intercepted from the line galloping video as the study object and applies the pyramid optical flow method to calculate the optical flow value of every point in the image. The isolated noise points are filtered by the closing operation of morphological filtering and the galloping transmission lines in the image are selected by the setting of area threshold. The image is processed by the multiple sub- pixels filtering to divide its vector field into the transverse and longitudinal real value images. The location of optical flow pixel is estimated in transverse and longitudinal direction respectively and the flow velocity is calculated. The direction and displacement of galloping movement are depicted for each point on the transmission line in the image for automatic conductor galloping monitoring. The method overcomes the deficiency of video monitoring that purely depends on staff to watching the status of conductor. Case analysis shows that,the galloping transmission lines in the image are properly identified and the direction and displacement of each point on the galloping transmission line are accurately calculated.

关 键 词：光流 输电 导线 舞动 监测 图像分析 

分 类 号：TM75[电气工程—电力系统及自动化]