城市居民出行O/D时空分布特征的轨迹数据提取

Trajectory data extract of O/D temporal and spatial distribution of urban residents

为了利用轨迹数据提取城市居民出行特征,该文首先从出租车全球定位系统(GPS)轨迹数据中提取上下客事件,利用数理统计的方法对其时间分布特征进行研究;继而针对城市内数据不均匀分布的特性,提出了一种基于OPTICS聚类的改进方法,通过将原始聚类结果进行局部密度峰值的计算,实现了出租车上下客事件热点区域中心的进一步提取。最后,以成都市为实证对象,结合成都市电子地图及用地布局规划图探讨了居民出行时空特征与城市空间结构的相互关系,并对其成因做出了简要的分析,证明了该文所提出方法的有效性。实验结果表明:居民在工作日和休息日的出行特征呈现出明显的差异,且受到城市功能区域分化的影响。

In order to extract urban residents’travel characteristics from trajectory data,this paper firstly used taxi global positioning system(GPS)trajectory data to extract pick-up and drop-off events by the algorithm,and then used mathematical statistics method to analyze their characteristics of time distribution.Next,aiming at the characteristics of uneven distribution of urban data,an improvement based on OPTICS clustering was proposed.At the same time,based on the results of ordering points to identify the clustering structure(OPTICS)clustering method,the local density peak value of clustering results was calculated,which was designed to extracting the hot spot centers accurately in the situation of the inhomogeneity distribution of the trajectory data intra city.Finally,taking Chengdu as an object of the research,we discussed the correlation between the characteristics of citizens’travelling and the urban spatial structure,based on Chengdu electronic map and land layout planning map and analyzed the causes of this phenomenon,which showed the effectiveness of the proposed method in this thesis.The experimental results indicated that the distribution and the lifecycle of urban hot spots was closely related to urban functions in different regions.