基于改进Meanshift算法的刀闸实时状态的精准跟踪  

Accurate tracking of real-time status of knife-switch using the improved Meanshift algorithm

作　　者：台德群 翁士状[2] 罗希 李宇 孙长翔 TAI Dequn;WENG Shizhuang;LUO Xi;LI Yu;SUN Changxiang(State Grid Anhui Electric Power Co.,Ltd.,Tongling Power Supply Company,Tongling 244099,China;School of Electronic Information Engineering,Anhui University,Hefei 230601,China;Anhui Nanrui Jiyuan Power Grid Technology Co.,Ltd.,Hefei 230088,China)

机构地区：[1]国网安徽电力有限公司铜陵供电公司,安徽铜陵244099 [2]安徽大学电子与信息工程学院,安徽合肥230601 [3]安徽南瑞继远电网技术有限公司,安徽合肥230088

出　　处：《安徽大学学报（自然科学版）》2020年第1期51-56,共6页Journal of Anhui University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(3170123,61672032)。

摘　　要：以变电站中的双柱式水平刀闸为例,基于传统Meanshift算法和局部二元模式(LBP)编码增强结构特征,提出改进的Meanshift算法,用该算法解析刀闸工作视频信息,实现刀闸实时状态跟踪.通过对87个刀闸开合视频的跟踪实验可知,改进算法在X方向的跟踪精度提高了120%,Y轴方向的跟踪精度提高了800%.因此,该算法为刀闸状态的准确识别提供了新的手段.Taking the double-column horizontal knife-switch in the substation as an example,based on the traditional Meanshift algorithm and the local binary pattern(LBP)encoding enhancement structure,an improved Meanshift algorithm was proposed to analyze the working information of the knife-switch to realize its state tracking of real-time status.Through the tracking experiment of 87 knife-switch opening and closing videos,the tracking accuracy of the improved algorithm in the X axis direction was improved by 120%,and the tracking accuracy in the Y axis direction was improved by 800%.Therefore,the improved Meanshift algorithm provided a new means for accurate identification of the state of the knife-switch.

关 键 词：刀闸 状态跟踪 MEANSHIFT算法 局部二元模式编码 

分 类 号：TM712[电气工程—电力系统及自动化]