A road network data acquisition method, device, equipment and computer storage medium

Abstract

The present application relates to a road network data acquisition method, device, device and computer storage medium, the method comprising the following steps: acquiring a vehicle trajectory image in a target area; inputting the vehicle trajectory image into a pre-trained road network prediction model, To make the road network prediction model output the target road network image of the target area, wherein the road network prediction model is obtained by training the road network historical records, and the road network historical records include historical vehicle trajectory images and historical vehicle trajectory images corresponding to Road network image: use the target road network image to generate target road network data in the target area. This application uses the trained road network prediction model to generate more accurate target road network data, which improves the generation accuracy and recall rate of the target road network data, and can restore the electronic map more accurately, effectively resisting the existence of vehicle trajectory data. It solves the problem of low sampling rate and uneven spatial distribution of vehicle trajectory data in different regions.

Description

technical field [0001] The present application relates to the technical field of machine learning, and in particular to a method, device, equipment and computer storage medium for acquiring road network data. Background technique [0002] With the rapid growth of the economy, the transportation is becoming more and more convenient, and there are new roads built every day across the country. In order to maintain the update of the map, the traditional method needs to manually collect road network data across the country, which consumes a lot of manpower and financial resources. With the process of digitalization and informatization, the cost of obtaining vehicle trajectory data is gradually decreasing. The trajectory data can reflect the connection mode from region to region. By aggregating a large number of trajectory data of different vehicles, it is possible to infer the road network structure information behind the trajectory data. Helps to automatically add new road segme...

AI Technical Summary

Problems solved by technology

[0003] The traditional method is usually to cluster the trajectory points of all vehicles to generate a road network, and to cluster the sub-trajectories of different vehicles, but the above method has the following defects: the road section with a small number of passes is unable to Generate a road network, there is a problem of low recall rate, and the sub-trajectory data has a large amount of calculation, and the clustering accuracy is poor, and many redundant and false road sections may be generated.

Images