Intersection broadening lane detection method based on gradient lifting decision tree

Abstract

The invention provides an intersection broadening lane detection method based on a gradient lifting decision tree. The method comprises the steps of performing data input, characteristic analysis andcharacteristic selection, and selecting the distribution width, distribution density, direction included angle of FCD, and speed of the floating vehicle data points on the cross section of the road; constructing a basic classifier, including that the basic classifier is constructed according to the running scene of the trajectory of the floating vehicle; performing the lane calculation based on the gradient lifting decision tree, including by means of the floating vehicle data, and referring to the basic classifier, selecting the classification number having the minimum information entropy andcorresponding to the optimal distribution state of the lane in the cross section when the FCD is divided into different lane numbers; performing broadening lane determination, including comparing thelane number changing conditions on the road sections corresponding to different research units on the same road, and determining whether the intersection has broadening lanes. According to the invention, the accuracy of intersection lane number determination is improved.

Description

technical field [0001] The invention belongs to the technical field of spatio-temporal trajectory big data, and relates to a detection method for widening lanes at intersections based on a gradient lifting decision tree. Background technique [0002] With the popularization of GPS positioning technology, more and more vehicles are equipped with GPS receivers. The trajectories of such vehicles almost cover the entire road network and road surface of the city [1,2]. These spatio-temporal GPS trajectory data are based on floating The road information acquisition technology of vehicle data provides a large amount of historical data. [0003] Using spatio-temporal GPS trajectory big data to extract urban road network and road lane information is a hot topic in current research. This section summarizes the relevant research progress in the calculation of the number of road lanes based on floating car trajectories and widening lanes at intersections from the perspective of data so...

AI Technical Summary

Problems solved by technology

Due to the limited amount of data, it cannot fully reflect the real situation of the intersection

At the same time, there are relevant regulations on the widening design of intersections in the existing domestic regulations, but there are problems of different standards and conflicts[1]

Existing research results have shown that although widening lanes can improve the traffic capacity of intersections to a certain extent, due to its special geometric conditions and operating characteristics, road traffic accidents such as scratches often occur[12,16]

Images