Traffic signs as a small target, its fast and accurate detection in the complex traffic environment has certain practical significance. This paper focuses on the lack of traffic sign detection accuracy in complex and changing traffic scenes, and improves the YOLOv7 algorithm in deep learning to improve the accuracy of the current existing algorithms and meet certain real-time performance. Experimental comparisons were conducted on the TT100k dataset, and the improved method improved the accuracy of the model by 1.04%, the recall by 2.37%, and the mean average precision by 2.52% compared with YOLOv7. The trial report has shown that the improved algorithm can validly enhance the precision of traffic sign detection and significantly improve the misdetection and omission of small.
Research on Deep Learning Based Traffic Sign Detection Methods
20.09.2024
2361371 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Traffic sign detection based on deep learning
SPIE | 2023
|Traffic sign detection using Deep learning techniques
IEEE | 2023
|Traffic Sign Detection and Recognition Using Deep Learning
Springer Verlag | 2022
|