A Comprehensive Study on Traffic Sign Detection in ITS

Abstract

Intelligent transportation system (ITS) is comprised of several component like Commuter information, Traffic management, vehicle info, public safety etc. To ensure the smooth flow of traffic, the recognition and apprehension of traffic signs are fundamental building blocks of the system. In real time scenario this system Driver Personal Assistant (DAP) can assist driver to travel in stress-free environment and which is one of the reason researchers are engrossed for expansion of such type algorithms in this direction. Due to various symbols of traffic sign images in various countries brought up by shifting perspectives, motion blur, lighting, and other parameters, accurate detection is a big challenging. Several techniques comprise of Artificial Intelligence, Machine learning and Deep Learning has been established in real time span for the enhancement of object detection and recognition of traffic sign. The content is potential for researchers’ association and has developed meaningful interest in several studies on optics -based TSDR (detection and Recognition of Traffic Sign). Three major component detection, Recognition, and classification combined in any algorithm make the system useful for deployment in real time. Variety of techniques had been communicated over a period. All-inclusive brief summary for traffic sign detection is presented in this paper. In addition to this exhaustive summary of algorithms, on detection, recognition and classification are investigated and grouped in tabular representation. An algorithm based on CNN has been implemented and shows better accuracy and less percentage loss in detection of GTSRB traffic sign data set.