密度聚类划分时间段的动态热度路网构建

Constructing dynamic hot road networks using time points density clustering

为减少构建路网计算量,快速获取交汇口,提出停留点扩展法。分析停留点空间分布特性,仅根据停留点的相关性进行扩展,减少轨迹点遍历,实现交汇口的快速获取。为准确将静态热度路网转化为动态热度路网,提出密度聚类划分时间段法。对轨迹点使用密度聚类,计算聚类结果中各个时间段的热度,作为边的动态热度,实现动态热度路网构建。实验结果表明,停留点扩展法能够减少计算量,快速获取交汇口;密度聚类划分时间段法可以根据不同时段轨迹点密度变化合理划分时间段,实现动态热度路网构建。

To reduce the computation costs in road networks construction,an enhanced stop points expansion approach was proposed to efficiently acquire intersections.Features of the distribution of stop points were analyzed,the expansion was judged according to the relevance of stop points,which reduced the points traversal and acquired intersections effectively.To transform a static hot road network into a dynamic hot road network,the time points density clustering was proposed.The points were clustered by density clustering,the hot value in each cluster was calculated as the dynamic hot value of edge,which completed the construction of dynamic hot road networks.Experimental results show that the proposed stop points expansion approach reduces the computation costs to enhance the efficiency of acquiring intersections.The time points density clustering partitions time properly according to the points density variation trend which can construct a dynamic hot road network.