一种基于轨迹特征划分的交通轨迹数据分析方法

An Traffic Trajectory Data Analysis Method Based on Trajectory Feature Division

当前大多交通轨迹数据的划分方法并未考虑空间分布的任意性,划分点较单一,从而导致分析效果并不理想。针对此问题,本文提出了多特征轨迹数据点结合数据空间划分方法,对提取采集到的海量交通轨迹的记录点,利用α-Shapes算法进行预处理并去噪,计算轨迹特征点,对这些特征点按空间接近度进行分组,之后按照位置关系进行Voronoi划分。该方法克服了轨迹数据划分时因空间分布任意性导致的划分效果不明显的缺点,有效地提高了轨迹数据分析效果。采用山东省淄博市出租车数据对本方法进行验证,结果证明该方法较传统数据划分方法在效果上优势明显,在轨迹数据去噪方面也做出了贡献。

Most of the current division methods of traffic trajectory data do not take into account the arbitrary nature of its spatial distribution,combined with the division of a single point,resulting in the analysis is not ideal. In order to solve this problem,a multifeature trajectory data point combined with data space division method is proposed.The recorded points of the massively-trafficked traffic trajectory are extracted,pre-processed and denoised by α-Shapes method,and the trajectory feature points are calculated. Click the spatial proximity to group,and then according to the location of the Voronoi division. The method overcomes the shortcomings that the division effect is not obvious due to the randomness of spatial distribution when the trajectory data is divided,and effectively improves the trajectory data analysis effect.This method is validated by the taxi data of Zibo city,Shandong province.The results of the distribution heat map after the data are divided show that this method is more effective than the traditional data classification method and also contributes to the de-noising of the trajectory data.