This study describes an application that uses a neural network approach to traffic sign recognition. Traffic sign recognition is an important part in the assessment of traffic situations by autonomous and intelligent vehicles. Although road signs are standardized in size and shape in every country, there can be difficulties in detecting and recognizing them in the video stream, so improving the accuracy of their recognition is an urgent task. To solve the problem of recognition of road signs the up-to-date real-time object detection system YOLOv5 was used. To train the neural network we used the data set consisting of 10 classes of approximately 200 images each. According to the results of system testing, the recognition accuracy was 72%.
Traffic Sign Recognition Application Using Yolov5 Architecture
2021-09-05
4499955 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Research and Implementation of an Embedded Traffic Sign Detection Model Using Improved YOLOV5
Springer Verlag | 2024
|TRAFFIC SIGN RECOGNITION DEVICE AND TRAFFIC SIGN RECOGNITION METHOD
Europäisches Patentamt | 2023
|