Method and device for obtaining vehicle queuing length based on image feature fusion

Abstract

The invention discloses a method and device for obtaining the vehicle queuing length based on image feature fusion, and relates to the field of intelligent transportation. The method comprises: obtaining continuous adjacent multi-frame image data collected by a target monitoring point, obtaining a feature mask matrix of a vehicle image based on a semantic segmentation network model, obtaining a feature coding matrix of a vehicle image in a first lane based on a lightweight image coding neural network model, performing channel splicing on the feature mask matrix and the feature coding matrix to obtain a spliced feature matrix, and obtaining an image fusion feature matrix based on a constructed lightweight feature fusion neural network model; and determining the vehicle queuing length on the first lane of the target monitoring point according to the queue state constructed by the lightweight neural network, a length prediction model and the image fusion feature matrix. The queuing length prediction accuracy can be effectively improved to a certain extent, and the prediction efficiency can be improved. The method is suitable for intelligent control scenes of traffic lights and reversible lanes.

Description

technical field [0001] The invention relates to the field of intelligent transportation, in particular to a method and a device for acquiring vehicle queuing length based on image feature fusion. Background technique [0002] Currently, known methods for determining vehicle queue length include: [0003] (1) Automatically detect the queue length of vehicles at the intersection based on the texture features of the empty lane area at the end of the queue, and use the maximum similarity registration method of the fixed background area window to solve the deviation caused by jitter. [0004] (2) Based on the equivalent queuing length model of a single-lane road section, the maximum equivalent queuing length model was established, and its traffic flow characteristics, time characteristics and spatial characteristics were analyzed by mathematical statistics methods, and the single space-time parameter was used to solve the maximum equivalent queuing length model by partial differe...

AI Technical Summary

Problems solved by technology

[0008] The inventor found in the process of realizing the present invention: for the above several methods, the background modeling of the actual area needs to be carried out in the schemes (1) and (5), and the adaptability is poor; the statistical method is used in the scheme (2), which is inconvenient The real-time queuing length is obtained, and the short-term error is relatively large; the corner feature error in scheme (3) is relatively large, and the effect is not good for poor weather; the adjacent frame difference method and connected domain search in scheme (4) introduce more Experience threshold, insufficient adaptability, affects the prediction accuracy of vehicle queuing length

Images