Traffic sign identification method and system based on coordinate attention

Abstract

The invention discloses a traffic sign identification method and system based on coordinate attention. The method comprises the following steps: receiving an image data frame of a front road in real time; performing pooling decomposition in the two-dimensional direction after performing cyclic down-sampling on the image data frame to obtain a plurality of decomposed feature image blocks; the position information of the decomposed feature image blocks is embedded into channel attention, splicing fusion of all stages is carried out according to the same dimension information in the feature image blocks, and a target feature image is obtained after convolution processing in the two-dimensional direction; and in response to a preset traffic sign map library, after obtaining the target feature image, classifying the target feature image by using the activation function, and obtaining traffic sign information in the image data frame. According to the method, the detection speed and the recognition precision are balanced, the feature extraction capability of the network is improved, the detection effect on the shielded target and the small target is improved, and the road traffic sign is detected in real time in a real scene.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method and system for identifying traffic signs based on coordinate attention. Background technique [0002] Traffic sign recognition is very important in the field of unmanned driving and assisted driving, but some traffic accidents are caused by the failure of drivers to accurately locate the location and meaning of traffic signs. Therefore, efficient traffic sign recognition technology can help. To reduce the occurrence of traffic accidents and improve the efficiency of road traffic. [0003] At present, the traffic sign recognition technology used in vehicle technology is mainly divided into detection algorithms based on traditional algorithms and detection algorithms based on deep learning. Among them, in the related research on traditional detection algorithms, Fleyeh H et al. used color segmentation based on AdaBoost binary classifier and cyclic Hough transfor...

AI Technical Summary

Problems solved by technology

[0005] The embodiment of the present application provides a traffic sign recognition method and system based on coordinate attention, which solves the problems in the prior art that due to the complex environment in the real road scene, traffic sign recognition is difficult to detect due to illumination, target occlusion, and small targets. , leading to low recognition accuracy, and the technical problems that the driving speed in real road scenes cannot be quickly recognized, and the purpose of accurately recognizing traffic signs in small targets through the traffic sign recognition model has been realized, and the balance between detection speed and recognition accuracy has been achieved. Performance, improve the feature extraction ability of the network and improve the detection effect of occluded targets and small targets, real-time detection of road traffic signs in real scenes

Images