Traffic sign target detection method based on deep learning
Jiachen Jiang, Jianan Yang, Jiankai Yin · 2021 International Conference on Computer Information Science and Artificial Intelligence (CISAI) · 2021
With the development of autonomous driving technology, traffic sign recognition technology requirements are further improved. This paper conducts training and testing on the public data set of traffic sign recognition and performs filtering and data enhancement on the original data set to strengthen the detection of small targets. Research on common target detection algorithms such as Faster R-CNN, SSD (Single Shot MultiBox Detector), RetinaNet, yolov3 (You Only Look Once), yolov5, and optimize their network parameters. Experimental results show that the test accuracy of Faster RCNN, RetinaNet, and YOLO algorithms all reach more than 98.2%. The SSD algorithm has relatively low test accuracy due to the shallow number of layers, but it performs well in training and detection speed.