基于GPS轨迹的城市拥堵区域挖掘与分析

摘要

With the development of world urbanization,and the rapid growth of motor vehicle,traffic congestion problem is increasingly serious and has become a hot research topic. Traditional congestion detection generally uses the video surveillance or sensors,although they can effectively reflect the real-time congestion,but can't mine the law of the urban congestion,even fail to analyze the congestion correlation between urban areas. In this paper,the congestion region mining algorithm is proposed based on DENCLUE,first preprocess-ing the urban taxi GPS data,calculation of the congestion point,then clustering them by DENCLUE to determine the congestion area. It is shown in the experiment that the algorithm can effectively find out congestion areas and grade them,getting the congestion state of the cit-y,reflection of urban congestion. In addition,Spearman rank correlation coefficient is used to calculate the flow of car correlation coeffi-cient between regions,analysis of the congestion causes and effects of severe congestion regions combined with the actual location.%随着世界城市化进程的发展，以及机动车保有量的飞速增长，交通拥堵问题日益严峻，城市道路交通拥堵问题已成为研究热点。传统的拥堵检测手段一般采用视频监控或传感器检测，虽然可以实时有效地反映拥堵状态，但无法挖掘城市拥堵规律，更无法有效分析拥堵原因和拥堵影响。文中提出基于DENCLUE的拥堵区域挖掘算法，对车辆GPS数据预处理，计算出拥堵点，然后对拥堵点进行DENCLUE聚类来确定拥堵区域。实验证明，该算法可以有效找出拥堵区域，并将拥堵划分等级，得到城市区域拥堵状态，反映城市拥堵情况。此外，使用斯皮尔曼等级相关系数计算区域间拥堵程度相关系数，结合实际地理位置分析拥堵区域的拥堵原因及影响。