首页 | 本学科首页   官方微博 | 高级检索  
     

煤矿视频分布式实时处理平台的设计与实现
引用本文:徐永刚,陈彬.煤矿视频分布式实时处理平台的设计与实现[J].煤矿机械,2021,42(1):164-167.
作者姓名:徐永刚  陈彬
作者单位:中国矿业大学信息与控制工程学院,江苏徐州221008;徐州市智能安全与应急协同工程研究中心,江苏徐州221008;中国矿业大学信息与控制工程学院,江苏徐州221008
基金项目:中国矿业大学“工业物联网与应急协同”创新团队资助计划(2020ZY002);国家自然科学基金面上项目(61971422;31871337)。
摘    要:实时智能化监控技术是煤矿安全生产的关键。针对煤矿井下监控视频处理实时性差、数据挖掘率低等问题,基于Spark、Kafka、HDFS等分布式框架设计并搭建了一个视频分布式实时处理平台,提出了一种多节点加权融合算法,改进了帧间相关视频的处理方式,加入关联分析模块,并进行了实验验证。实验结果表明,该平台在帧间相关和帧间无关的视频处理中均具有良好的实时处理能力,提高了煤矿视频监控的实时监测能力,对煤矿的安全生产具有重要的意义。

关 键 词:煤矿  监控  Spark  Kafka  分布式  实时处理

Design and Implementation of Video Distributed Real-time Processing Platform for Coal Mine
Xu Yonggang,Chen Bin.Design and Implementation of Video Distributed Real-time Processing Platform for Coal Mine[J].Coal Mine Machinery,2021,42(1):164-167.
Authors:Xu Yonggang  Chen Bin
Affiliation:(School of Information and Control Engineering,China University of Mining and Technology,Xuzhou 221008,China;Xuzhou Engineering Research Center of Intelligent Industry Safety and Emergency Collaboration,Xuzhou 221008,China)
Abstract:Real-time intelligent monitoring technology is the key to safe production in coal mines.Aiming at the problems of poor real-time processing of underground monitoring video in coal mines and low data mining rate,designed and built a video distributed real-time processing platform based on distributed frameworks such as Spark, Kafka, HDFS, and proposed a multi-node weighted fusion algorithm.Improved the processing method of inter-frame related videos, added the correlation analysis module,and carried out experimental verification. The experimental results show that this platform has good real-time processing capability in the video processing of inter-frame correlation and inter-frame irrelevance, which improves the real-time monitoring capability of video monitoring in coal mine, and is of great significance to the safety production of coal mine.
Keywords:coal mine  monitoring  Spark  Kafka  distributed  real-time processing
本文献已被 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号