大数据在监管中心智能化系统中的应用研究  

Research on the application of big data in the intelligent system of supervision center

作　　者：杨彦青[1] 宋星[1] YANG Yanqing;SONG Xing(Taizhou Vocational and Technical College,Zhejiang Taizhou 318000,China)

机构地区：[1]台州职业技术学院,浙江台州318000

出　　处：《自动化与仪器仪表》2021年第1期34-37,42,共5页Automation & Instrumentation

基　　金：浙江省教育厅高等学校访问工程师校企合作项目“大数据在监管中心智能化系统中的应用研究”(No.FG2019165)。

摘　　要：监管视频数据的剧增,导致现有系统无法满足社会的监管需求,为此提出大数据在监管中心智能化系统中的应用研究。此研究着重分析大数据技术在监管中心智能化系统—智能视频分析模块的应用,基于大数据技术设计目标行为分析软件,包括监管视频数据处理层与应用层。监管视频数据处理层采用数据整理技术—Retinex算法降噪、增强视频图像,监管视频数据应用层以处理层得到的监管视频数据为基础,利用光流法识别视频图像中的目标,通过数据分析技术—Mean-Shift算法跟踪目标,提取目标行为特征,依据MF推理方法检测目标异常行为,实现了基于大数据的目标行为分析软件的运行。测试结果显示:应用大数据的监管中心智能化系统,目标异常行为检测准确性大于66.12%,目标异常行为报警时间小于限值9.45 s,充分表明大数据在监管中心智能化系统中应用的有效性。With the rapid increase of surveillance video data,the existing system can not meet the regulatory needs of the society.Therefore,the application of big data in the intelligent system of the supervision center is proposed.This research focuses on analyzing the application of big data technology in the intelligent system of Supervision Center-intelligent video analysis module.Based on big data technology,the target behavior analysis software is designed,including supervision video data processing layer and application layer.In the supervisory video data processing layer,Retinex algorithm is used to reduce noise and enhance video image.In the supervisory video data application layer,based on the supervisory video data obtained by the processing layer,optical flow method is used to identify the target in the video image.Through data analysis technology,mean shift algorithm is used to track the target,extract the behavior characteristics of the target,and detect the abnormal line of the target according to MF reasoning method,in order to achieve the goal behavior analysis software based on big data.The test results show that the accuracy of detection of target abnormal behavior is more than 66.12%,and the alarm time of target abnormal behavior is less than the limit value of 9.45 s,which fully shows the effectiveness of big data application in the supervision center intelligent system.

关 键 词：大数据 监管中心 智能化 异常行为 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]