In day to day lives, road safety has always been a major concern. A lot of road accidents occur in India. Many of them are fatal. Hence, in this paper, an automatic traffic sign recognition system is proposed which recognises the traffic signs and informs the same to the driver. Two cameras are installed on the vehicle, focussing on two sides of the road. The video feeds from both cameras are sent to the recognition system continuously. The system converts the videos into frames. The sign detected is matched with the database image using Normalized Cross Correlation technique. Two levels of parallelism are adopted to perform the operations in a faster manner. First level is at Frame conversion and Detection and second at the recognition stage. If the image is matched, then it is informed to the driver. The method used is orientation dependent. Hence, signs like turn right and turn left are not confused with each other. Also, the reduction in execution time obtained by parallelism provides drivers more time to respond to the signs.
Automatic traffic sign recognition system using fast normalized cross correlation and parallel processing
01.12.2017
597194 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Automatic video processing for traffic sign recognition
Tema Archiv | 2009
|Traffic Sign Recognition for Automatic Vehicles
IEEE | 2024
|