Traffic sign recognition model training method and system

Abstract

The invention provides a traffic sign recognition model training method and system, and the method comprises the following steps: obtaining traffic sign images and a label corresponding to each traffic sign image, and dividing the traffic sign images into a training set and a test set; obtaining a network model based on YOLOV5, and loading and setting the network model; setting coefficients and hyper-parameters of classification loss of the network model according to categories of the training set and the test set; carrying out warm-up training on the network model by adopting mixed precision training, calculating the sum of three losses, and carrying out back propagation to carry out gradient amplification; and carrying out learning rate attenuation on the network model after warm-up training, and then storing the network model and the weight after the test set network model is tested to generate a traffic sign image model. According to the invention, the traffic signs can be effectively detected and classified in real time.

Description

technical field [0001] The invention relates to traffic sign recognition, in particular to a traffic sign recognition model training method and system. Background technique [0002] According to statistics, the proportion of traffic accidents caused by illegal driving of motor vehicle drivers in my country has reached more than 90%, part of the reason is that drivers cannot timely process the instructions of road traffic signs in visual information. Performance is very important for improving the assisted driving system, but the traditional traffic sign recognition speed is slow, the detection effect is not ideal, and it is difficult to be used in actual scenarios. [0003] As convolutional neural networks and target detection algorithms are widely used to deal with image recognition and classification problems, and the automotive industry pays more and more attention to safety performance, deep learning has been widely used in research in this area. Chen Mingsong and others...

AI Technical Summary

Problems solved by technology

The traditional traffic sign recognition method is easily affected by light, occlusion, and the target is too small, it is difficult to detect traffic signs quickly and accurately, and there are potential safety hazards. An effective method is needed for real-time detection and classification

Images