基于监控视频的变电站运动目标提取  

Moving Target Extraction of Substation Based on Surveillance Video

作　　者：胡旦华 刘燕萍 张宇晴 HU Danhua;LIU Yanping;ZHANG Yuqing(Power China Fujiaa Electric Power Engineering Co.,Ltd.,350003,Fuzhou,PRC;Tongji University Zhejiang Collece,314051,Jiaxing,Zhejiang,PRC;School of Mathematics and Information,Fujian Normal University,350007,Fuzhou,PRC)

机构地区：[1]中国电建集团福建省电力勘测设计院有限公司,福州350003 [2]同济大学浙江学院,浙江嘉兴314051 [3]福建师范大学数学与信息学院,福州350007

出　　处：《江西科学》2020年第6期896-902,共7页Jiangxi Science

基　　金：嘉兴市科技计划项目(2019AY11017)。

摘　　要：虚拟场景三维模型的建立对数字孪生技术的管理和应用具有重要意义,而现实视频监控与虚拟场景模型结合起来,实现虚拟三维场景中的实况监控投影,将便于人们更加真实地再现现实场景,更加直观地从模型中提取所需的目标信息,更加方便有效地实现虚拟现实管理。详细讲述了变电站虚拟三维场景中链接投影真实监控视频的方法,同时针对输入的视频信息,采用改进的混合高斯背景模型实现了运动对象的实时提取,同时和帧间差分法进行比较分析,证实了改进的混合高斯背景模型在运动目标提取中的准确性和稳健性。该方法的研究对变电站的安全运维具有重要的应用价值。The establishment of three-dimensional model of virtual scene is of great significance to the management and application of digital twins technology.The combination of real-time video monitoring and virtual scene model to realize the real-time monitoring projection in virtual three-dimensional scene will facilitate people to reappear the real scene more realistically,extract the required target information from the model more intuitively,moreover,to realize virtual reality more conveniently and effectively Administration.In this paper,the method of linking and projecting the real monitoring video in the virtual three-dimensional scene of substation is described in detail.According to the input video information,the improved Gaussian mixture background model is used to realize the real-time extraction of moving objects.At the same time,compared with the inter frame difference method,the accuracy and robustness of the improved Gaussian mixture background model in moving object extraction are verified.The research of this method has important application value for the safe operation and maintenance of substation.

关 键 词：虚拟场景 数字孪生 监控视频 混合高斯背景模型 运动目标提取 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]