Design of Traffic Sign Detection and Recognition Algorithm Based on Template Matching

摘要

In the field of automobile unmanned driving and assisted driving, the algorithms of traffic sign detection and recognition are extremely important, and the performance of these algorithms directly affects the driving safety of motor vehicles in road traffic. Especially today, it is necessary to ensure the accuracy, real-time, stability and anti-interference of traffic sign recognition, which have always been the focus of research. The new algorithm mainly performs the following tasks for the detection and recognition of traffic signs: first, the image is preprocessed to improve the image quality and reduce the amount of data. The image preprocessing mainly includes grayscale, image enhancement, image restoration and binarization; Secondly, we need to perform edge detection on the traffic sign image, and use color segmentation to roughly segment the image of interest; next, we perform morphological processing on the part obtained by the rough segmentation, and use shape segmentation to eliminate unnecessary interference, so that we can obtain Interested part; Finally, the final cut traffic signs are compared with the template library for recognition. The simulation results show that the algorithm in this paper can effectively recognize the traffic sign images, which can greatly improve the recognition rate and has extremely broad development prospects and research value.