Automatic Road Sign Detection, Recognition and Notification System

Abstract

As, road accidents are becoming very frequent these days, the community comes up with numerous rationales for this crisis. However, it is globally accepted that pay no attention to the traffic /road signs which have been placed either sides of the road is one of the major causes of mounting traffic jam and road accidents. In the mean time it is reasonable that the driver might not get recognized the placed road signs due to caring of passing cars, un-expected happening is in the road sides, vehicle speed and other orientations, etc. To overcome this concern, a driver assistant system that is embedded in a car is very much helpful to the driver for safe and comfortable driving. The proposed system (Automatic Road Sign Detection, Recognition and Notification System), takes care of that difficulty while automatically detecting and notifying the driver which sign has recognized. This improves the comfort and safe driving till the destined the journey. Since, different color schemes have been used in sign boards in order to provide different information the proposed system includes the color segmentations as well. Usually, yellow color used in traffic signs conveys general warnings and blue is used to provide services and guidance. In the mean time, the color red is used to indicate a prohibition or a stop. The RED colored road signs have been used to detect in this study as a proof of concept (POC) and a pre-defined text that is bound with the recognized road sign is delivered as a voice message through a speech synthesis. The positive detection percentage of red colored traffic signs by the proposed system is 90.91% and a favorable percentage amount (70.00%) of traffic signs in the entire domain of traffic signs have been recognized in gray scale by this system. Meanwhile, 80% of rotated background images are positively recognized by the concept used in this study.