Traffic Sign Recognition Based on YOLOv5 Network Structure
Tong Hu, Jun Song · 2022
Aiming at the problems of low recognition accuracy, slow rate, missed detection and false detection of traditional target detection algorithm in traffic sign recognition system, The deep learning technology can extract the feature information of traffic sign automatically, so a traffic sign target detection algorithm based on YOLOV5 network structure is proposed. The attention mechanism is added to enhance the effective features of the detected objects, suppress the interference features, and improve the performance of traffic sign detection in the complex background. The results show that survey data for training and testing were collected from China TT100K, and the optimized YOLOv5 algorithm has a good detection effect.