Traffic sign recognition model training method based on YOLOv4 and system

Abstract

The invention provides a traffic sign recognition model training method based on YOLOv4, and the method comprises the following steps: obtaining a traffic sign data set, and dividing the traffic sign data set into a training set and a test set; obtaining a YOLOv4 network structure, and replacing a BiFPN network with a PANet path aggregation network of a backbone network in the YOLOv4 network structure to generate a target training model; and training the target training model through the training set, and testing the trained target training model through the test set to generate the traffic sign recognition model. The backbone network has a better detection effect, the problem of unbalanced data distribution is relieved, and the feature extraction capability of the network is enhanced.

Description

technical field [0001] The present invention relates to an automatic driving system, in particular to a YOLOv4-based traffic sign recognition model training method and system. Background technique [0002] In recent years, self-driving cars have attracted widespread attention from the society due to their advantages of safety and efficiency. Google's driverless fleet has been tested in several states, but many traffic accidents occurred during the test run. It can be seen that the detection and recognition of road traffic signs is of great significance for improving the safety and reliability of automatic driving systems. [0003] With the rapid development of target detection algorithms and the increasing emphasis on safety performance in the field of unmanned driving, deep learning has been widely used in the research of road traffic signs. Liu Sheng et al. used a traffic sign recognition network based on multi-scale feature fusion to make full use of multi-scale features...

AI Technical Summary

Problems solved by technology

[0004] The current unmanned driving technology has great potential safety hazards. The traffic sign recognition technology equipped with the roof sensor is easily affected by light, occlusion and weather. The detection effect of road traffic signs in different environments is quite different, so lead to traffic accidents, so an effective method is needed to improve the detection accuracy of road traffic signs

Images