摘要
针对民用机场航站楼群体性事件预警问题,提出了基于视频监控图像分析的非正常人群外部聚集特征识别和预警方法。该方法通过提取图像中的人员位置信息,采用核密度空间聚类算法,人群聚集的空间和运动特征,建立航站楼旅客群体性事件案例库,基于模糊-粗糙集方法分析群体性事件中非正常聚集人群特征,结合相应的机场运行信息,建立相应的预警规则。理论分析和历史数据案例测试结果表明,该方法针对航站楼群体性事件的预警准确率明显高于单一的人群密度分级预警方法。
Outside accumulation feature recognition and pre-warning method was proposed to solve terminal mass events warning.With the assistance of passengers＇ locations and kernel density clustering algorithm,passengers＇ spatial characteristics and motion feature were clustered.On groundwork above,a case base of mass events in terminals was built.Adopting fuzzy-rough set algorithm,the base evaluates the features of the abnormal accumulating crowds in terminal mass events.Besides,the rules of mass events pre-warning was developed according to airport operation information.Theoretical analyses and tests on history operation data demonstrate a higher accuracy of terminal mass disturbance early warning,compared with the methods only based on crowd density classifications.
出处
《解放军理工大学学报（自然科学版）》
EI
北大核心
2013年第5期501-506,共6页
Journal of PLA University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金重大研究计划资助项目(91024024)
江苏省自然科学基金资助项目(SBK201241637)
关键词
民用航空
群体性事件
视频监控
人群聚集特征
预警
civil aviation
mass events
video surveillance
accumulation features of crowds
pre-warning
