图像识别技术在电力设备监测中的应用  

Application of image recognition technology in electrical equipment on-line monitoring

作　　者：张浩[1] 王玮[1] 徐丽杰[1] 秦欢[2] 刘明[3] 

机构地区：[1]北京交通大学电气工程学院,北京100044 [2]华北电力大学电气与电子工程学院,北京102206 [3]白山供电公司,吉林白山134300

出　　处：《电力系统保护与控制》2010年第6期88-91,共4页Power System Protection and Control

摘　　要：为减少传统基于灰度的匹配算法的计算量,采用图像形状特征分析技术,提出了新的模板匹配算法。该算法取隔行隔列的数据进行粗略匹配,在此基础上进行精确匹配,大大减少了计算的数据量。将该算法应用于电力变压器的图像识别中,实验结果表明该匹配算法在速度上具有较大的优势,能更快地对电力变压器进行定位识别。论述了远程数字视频监控与图像识别技术在电力系统中的应用前景。In order to reduce the amount of calculation of the traditional matching algorithm which is based on grayscale,the shape characteristics of image analysis technology is used and a new template matching algorithm is proposed.The algorithm first checks out the following interlacing data and roughly matches them,and then does precise matching.Therefore,it can greatly reduce the mount of data calculation.The algorithm which is applied in the image recognition of power transformer has an advantage of speed of calculation,and it can more quickly confirm the position of the power transformer.This paper also points out the prospect of applying remote digital video monitoring and image recognition technology in power systems.

关 键 词：电力设备 模板匹配 视频监控 图像识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TM73[自动化与计算机技术—计算机科学与技术]