基于多层聚类的自适应人群分群检测算法
|
摘要
在群体检测的仿真研究中,针对群体运动的特征,提出一种通过多层聚类的自适应分群检测算法。首先,在获得检测视频的前景区域后提取KLT特征点;再通过分析特征点的运动特征,分别以特征点的距离和加速度方向为多层聚类的输入;最后,引入聚类中心的社会力模型以消除多层聚类出现的分类误差。实验结果显示,相比于其他分群检测算法,所提算法降低了特征点的错误率,同时提高了分群数量的准确度。
In the simulation study of group detection, aiming at the characteristics of group motion, an adaptive cluster detection algorithm through multi-layer clustering was proposed. First, the KLT feature points were extracted after obtaining the foreground area of the detected video. Then, by analyzing the motion characteristics of the feature points, the distances of the feature points and the acceleration direction were used to be the input of the multi-layer cluster respectively to complete the filtering. Finally, the social force model was introduced between the cluster centers to eliminate the classification error of multi-layer clustering. The experiment results showed that the error rate of feature points was reduced and the accuracy of clustering was improved through the proposed algorithm compared with other cluster detection algorithms.
In the simulation study of group detection, aiming at the characteristics of group motion, an adaptive cluster detection algorithm through multi-layer clustering was proposed. First, the KLT feature points were extracted after obtaining the foreground area of the detected video. Then, by analyzing the motion characteristics of the feature points, the distances of the feature points and the acceleration direction were used to be the input of the multi-layer cluster respectively to complete the filtering. Finally, the social force model was introduced between the cluster centers to eliminate the classification error of multi-layer clustering. The experiment results showed that the error rate of feature points was reduced and the accuracy of clustering was improved through the proposed algorithm compared with other cluster detection algorithms.
引文
[1]QIU F,HU X.Modeling group structures in pedestrian crowd simulation[J].Simulation Modelling Practice&Theory,2010,18(2):190-205.
[2]GE W,COLLINS R T,RUBACK B.Automatically detecting the small group structure of a crowd[J].Workshop on Applications of Computer Vision,2009,12(7):1-8.
[3]WANG W,LIN W,CHEN Y,et al.Finding coherent motions and semantic regions in crowd scenes:a diffusion and clustering approach[J].Computer Vision,2014,9(2):756-771.
[4]B.ZHOU,X.TANG,X.WANG.Coherent filtering:detecting coherent motions from crowd clutters[J].Computer Vision,2012,10(1):857-871.
[5]RODRIGUEZ A,LAIO A.Clustering by fast search and find of density peaks[J].Science,2014,344(6191):1492-1496.
[6]邵洁,赵倩.基于多层自适应聚类模型的密集人群分群检测算法[J].上海电力学院学报,2017,33(1):91-96.
[7]HELBING D,MOLNAR P.Social force model for pedestrian dynamics[J].Physical Review E,1998,51(5):4282-4286.
[8]邵洁.基于改进社会力模型的密集场景目标方向预测[J].上海电力学院学报,2018,34(1):81-84.
[2]GE W,COLLINS R T,RUBACK B.Automatically detecting the small group structure of a crowd[J].Workshop on Applications of Computer Vision,2009,12(7):1-8.
[3]WANG W,LIN W,CHEN Y,et al.Finding coherent motions and semantic regions in crowd scenes:a diffusion and clustering approach[J].Computer Vision,2014,9(2):756-771.
[4]B.ZHOU,X.TANG,X.WANG.Coherent filtering:detecting coherent motions from crowd clutters[J].Computer Vision,2012,10(1):857-871.
[5]RODRIGUEZ A,LAIO A.Clustering by fast search and find of density peaks[J].Science,2014,344(6191):1492-1496.
[6]邵洁,赵倩.基于多层自适应聚类模型的密集人群分群检测算法[J].上海电力学院学报,2017,33(1):91-96.
[7]HELBING D,MOLNAR P.Social force model for pedestrian dynamics[J].Physical Review E,1998,51(5):4282-4286.
[8]邵洁.基于改进社会力模型的密集场景目标方向预测[J].上海电力学院学报,2018,34(1):81-84.