Traffic sign recognition by using chain code

Abstract

The Traffic signs are considered as Traffic Safety tools, Because of their role in the organization of traffic and vehicles to insure the safety of the passengers, pedestrians and the structures of traffic signs are very important devices which help the drivers to drive in safety and adequate manner. In this research Traffic sign detection performed in two stages: The first stage include the traffic sign detection and extraction from the road image scene, depending on the color features of the sign, the red color of the image was taken by using RGB color space system and applying threshold method, in which, for each layer specific threshold was applied. Considering the information of the external shape to recognize the shape geometry type of external frame by using Chain Code. While the second stage include traffic sign classification depending on inner contents of the sign, depending on the number of objects found in the inside part of the sign. Then, the Chain Code was used to recognize the boundary of the inner content of the sign. The research applied on a group of images with (.bmp, .jpg) extensions and with various sizes. The distinction percentage was (99%), the database included 30 images; 15 of them are warning and the other 15 are regulatory.