摘要
在利用检测移动设备进行人流量估测的方法中,存在着由于单人携带多个终端使总人数重复计算的问题。提出了一种根据定位轨迹之间的相关关系识别重复轨迹的算法,通过此方法可以去除人数的重复统计,提高总人流量估算的准确度。首先,针对多种典型轨迹的情况进行分析,得出3种基本类型的轨迹：交叉、并行和缠绕;其次,基于相关性检测的数学理论方法,提出了一种根据距离和空间相关性联合检测识别重复轨迹的算法;最后,在人流轨迹数据上对算法进行了仿真和性能分析。结果表明此算法能够检测出单人所携带多终端所产生的高相关性重复轨迹,显著地减少了因为重复轨迹产生的人流量估算误差。
In the method of estimating crowd density and pedestrian flow by sensing the mobile devices,there is a problem that a number of individuals carrying multiple terminals which leads to a repeated count of total number ofpeople. In this paper,an algorithm for identifying repetitive trajectories based on the correlation between locating trajectories is proposed. Through this method,the repeated statistics of the number of people can be removed to improve the accuracy of the total population flow estimation. Firstly,an analysis of various trajectories and three basic types of trajectories： crossover,parallelism and winding,are proposed. After that,based on the mathematical theory of correlation detection,an algorithm for recognizing repetitive trajectories using joint detection of distance and spatial correlation is proposed. Finally,simulation is carried out base onpedestrian flow trajectories data and performance of the algorithm is also analyzed,the result shows that this algorithm can detect the high correlation repetition trajectory generated by the multi-terminal carried by an individual,which greatly reduces the estimation error of the pedestrian flow due to the repeated trajectory problem.
作者
李进
李强
王营冠
Li Jin;Li Qiang;Wang Yingguan(University of Chinese Academy of Sciences, Beijing 100049, China;School of Information Science and Technology, ShanghaiTech University, Shanghai 201210, China;Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Science, Shanghai 200050, China;Shanghai Internet of Things Co. ,Ltd. , Shanghai 200050, China)
出处
《电子测量技术》
2018年第2期77-82,共6页
Electronic Measurement Technology
基金
上海张江国家自主创新示范区专项发展资金重点项目(201505-JD-C104-034)
中国科学院科技服务网络计划(KFJ-SW-STS-155)项目资助
关键词
相关性检验
统计显著性
室内定位
人数估计
correlation test
statistic significance
indoor positioning
population estimation
