论文部分内容阅读
针对航站楼旅客群体性事件的智能化预警问题,在民用机场航站楼内的现有视频监控系统的基础上,提出旅客群体性事件预警监控系统架构和关键技术。针对航站楼内的视频监控系统的视频数据进行预处理;基于空间聚类算法,从图像中提取人群聚集特征;针对目标监控区域的历史视频和突发事件数据进行统计分析,研究正常和非正常两种情况下的人群分布特征,并建立相应的预警规则库;根据预警规则,在人群聚集特征识别的基础上,对非正常目标密集人群进行识别。该系统可根据航站楼旅客群集性事件的时空特性,针对不同事件、不同时间、不同地点采用不同的规则进行预警,提高了预警的可靠性。
Based on the existing video surveillance system in the terminal of civil airport, this paper puts forward the structure and key technologies of early warning and monitoring system for passenger mass incidents. Based on the spatial clustering algorithm, the feature of crowd aggregation is extracted from the image; the historical video and emergency data of the target monitoring area are statistically analyzed to study the normal and non-normal Normal distribution of people under two conditions and the establishment of the corresponding precautionary rule base; based on the precautionary rule, based on the crowd gathered feature identification, based on the recognition of non-normal target intensive crowd. According to the spatiotemporal characteristics of the cluster incident of terminal passengers, this system can use different rules for different events, different time and different locations to make early warning and improve the reliability of early warning.