轨迹数据的多特征融合及检测方法

Multi-feature Fusion and Detection Method of Trajectory Data

传统轨迹检测方法中的轨迹相似度仅从位置向量进行度量,忽略了轨迹数据的速度和时间特征,这导致轨迹检测结果无法全面反映实际状况,降低了检测结果的有效性.针对上述问题,提出一种面向多个特征向量的轨迹数据相似性度量及检测方法.该方法首先将轨迹数据映射到图模型描述的轨迹图中,每条轨迹是轨迹图的一个节点;针对各节点的速度、时间和空间特征,给出了适用其度量的三个核函数,通过加权求和实现三个特征向量的融合;然后采用每个节点的特征融合值来构建轨迹数据的相似矩阵及其对应的Laplacian矩阵,以此实现轨迹数据的相似性度量;最后,运用K-means聚类方法对轨迹图进行分割,通过对的图模型节点的划分来实现特征相似的轨迹数据划分到相同的类.在实验中,采用出租车和飓风数据,分别对算法的效率和准确性进行检验,实验结果显示本文提出算法是合理有效的.

In the traditional trajectory detection method,the trajectory similarity is only measured from the position feature,and the speed and time features of the trajectory dataare ignored,which leads to the fact that the trajectory detection results cannot fully reflect the actual situation and reduce the effectiveness of the detection results.In order to solve the above problems,we propose a method of trajectory data similarity measurement and trajectory detection for multiple features.Firstly,the trajectory data is mapped to the trajectory map described by the graph model,and each trajectory is a node of the trajectory map.According to the speed,time and position features of each node,three kernel functions for its measurement are given,and the fusion of the three eigenvectors is realized by weighted summation.Then,the similarity matrix of the trajectory data and its fusion value are constructed.The corresponding Laplacian matrix is used to measure the similarity of trajectory data.Finally,K-means clustering method is used to segment the track map,and the track data with similar characteristics can be divided into the same category by dividing the nodes of the graph model.In the experiment,taxi and hurricane data are used to test the efficiency and accuracy of the algorithm.The experimental results show that the algorithm proposed in this paper is reasonable and effective.