Road congestion recognition visualization method based on DBSCAN+

Abstract

The invention discloses a road congestion recognition visualization method based on DBSCAN+. According to the method, first, a large amount of uploaded data of OBD vehicle terminals of floating cars in a city is preprocessed and cleaned; slow travel GPS track point data in a unit time period is extracted according to instantaneous velocities of the floating cars; DBSCAN+ algorithm parameters are set, and initial class cluster blocks in a slow travel region are acquired through parallel clustering of extracted data points; by calculating surface distances among all the data points in each classcluster block respectively, the two farthest points in each class cluster block are found out, and line segments are fitted; map-matching deviation rectification is performed on the fitted line segments according to a topological relation among all road segments in an actual road network; and last, by calculating travel distances of different vehicles in the unit time period in each class clusterblock respectively and calculating average travel distances synthetically, the congestion degrees of all the class cluster blocks are judged and expressed in different colors for differentiation andvisualization. The method can adapt to large-scale city taxi OBD terminal GPS track data, brings convenience to real-time recognition of city road congestion through taxi running conditions and achieves a good visualization effect.

Description

technical field [0001] The invention relates to the field of traffic big data, in particular to a road congestion identification and visualization method based on DBSCAN+ (Density-BasedSpatial Clustering of Applications with Noise Plus). Background technique [0002] With the increasingly serious problem of urban traffic congestion, establishing an effective road congestion identification system and accurately identifying traffic congestion sections in cities has become a current research direction. [0003] In the existing traffic jam detection technology, the traffic jam detection methods can be divided into the following categories according to the different processing methods of traffic data and the selected eigenvalues: (1) Based on the vehicle speed: some vehicles are directly detected by vehicle GPS The obtained vehicle speed data can be used directly after processing, or can be calculated from the time and mileage data of the vehicle. Xu L et al. used floating car d...

AI Technical Summary

Problems solved by technology

[0004] These commonly used traffic congestion detection methods are based on the idea of ​​extracting effective data from traffic data processing. In many cases, because traffic congestion detection does not have accurate results as a criterion, research in this field is basically unsupervised learning, and the data cannot be divided. Form a training set and a test set to verify the experimental results, and there are many ways to judge the results, so there is no unified definition of traffic congestion

Images