摘要
常规系统聚类视频数据时,忽略了结构信息的相似度,导致远距离传输数据的丢包率较高、时效性较差,针对这一问题,基于大数据标签技术,设计了输电线路鸟害视频网络实时监控系统。以视频采集设备、视频服务器、监控终端组成系统总体架构,压缩视频图像为视频码流,传输至监控端;预处理视频码流图像数据,聚类关联输电线路鸟害数据,实时显示鸟害监控信息。实验结果表明,该系统的平均丢包率为3.55%,平均帧率为9908fbps,平均时延为5.51ms,降低了丢包率,保障了时效性。
When clustering video data in conventional system,the similarity of structure information is ignored,which leads to little association of clustering data cluster,high packet loss rate and poor timeliness of long-distance transmission of data in cluster.In order to solve this problem,a real-time monitoring system of transmission line bird harm video network is designed based on big data label technology.The overall architecture of the system is composed of video acquisition equipment,video server and monitoring terminal.The compressed video image is transformed into video stream and transmitted to the monitoring terminal;The video stream image data is preproccessed,and the bird damage data of transmission line is clustered,and the alarm label is put on the video frame of bird harm to display the real-time monitoring information of bird harm.The experimental results show that the average packet loss rate of the system is 3.55%,the average frame rate is 9908 fbps,and the average time delay is 5.51ms,which can reduce the packet loss rate and guaranteed the timeless.
作者
石小帅
黄擎
张天坤
张江超
SHI Xiao-shuai;HUANG Qing;ZHANG Tian-kun;ZHANG Jiang-chao(State Grid Yili River Power Supply Co.,Ltd.,Yining 835000,Xinjiang Uygur Autonomous Region,China)
出处
《信息技术》
2022年第11期149-154,共6页
Information Technology
关键词
采集设备
编码器
视频信号
滤波去噪
标签聚类
acquisition equipment
the encoder
video signal
filtering and denoising
label clustering
